> Rewrite a bundler in Rust and you haven't only made it faster. You've also shrunk the pool of JavaScript developers who can maintain it. The new tool still looks like a duck and quacks like a duck, but it's a different beast altogether. Its internals retreat behind a black box that fewer people hold the keys to. The source may still be open, but the door to contributions is closing.
Alternatively, there's a pool of JS developers who shouldn't be maintaining critical infrastructure to begin with.
It's not a black box, those codebases are usually open and the only thing holding you or anyone back is learning anything outside of a small pond of JavaScript.
Write non-browser-things in fast languages. It is not a complicated concept - even less so in an era where stuff is getting written for you.
Absolutely agree. If a JS developer doesn't know Rust, Go, Zig, or something, then they probably aren't going to do a good job at maintaining critical infrastructure, even if it were written in JS.
Why are people writing "critical" infrastructure in JS anyway, it is the wrong tool for the job.
Why are people writing critical infrastructure in [any language]?
In reality, the tools this article is referring to were written in javascript because they could be and it was the best tool for the job (according to the people who matter: the people who did the work).
If someone wants to rewrite them for speed and/or to chase the next shiny language, that's fine with me, but let's not kid ourselves that there was something wrong with using javascript in the first place.
Using the right abstractions (including the right number of abstractions) and the right data structures and algorithms is always going to trump the constants language choice can optimize.
> Using the right abstractions (including the right number of abstractions) and the right data structures and algorithms is always going to trump the constants language choice can optimize.
I was with you until this paragraph, which is straight up incorrect. Language choice absolutely matters if your goal is performance.
The most obvious counterexample is that the TypeScript team pretty much did a direct port from TS to Golang, without significantly rewriting their core algorithms or changing their data structures, and it sped up the TS compiler by 10X, which is an enormous performance improvement that would be silly to dismiss as just a "constant." This is because JS is a bad choice of language for implementing a fast compiler, primarily due to its poor support for multithreading.
When did the love of learning disappear from this field? I remember when everyone i talked to was learning something new just to better themselves?
> When did the love of learning disappear from this field?
Like 10 years ago? When you could get a 400k/yr job after a 3 month JS boot camp. The market got flooded with people who don’t care about making good stuff. Now it’s the “AI has commodified intelligence so why learn anything” crowd
> When you could get a 400k/yr job after a 3 month JS boot camp.
I get the point but these numbers are hyperbole and do not happen in real life.
Or just write faster JS using it's lower-level language features. Nothing about whatever collection of build tools the JS ecosystem uses "requires" rust/zig/c++/go for performance. It's a closed text transform problem.
The native languages are easier to get decent performance for sure but, braking the entire ecosystem and a generation of future contributors, for what should realistically be single or low double digit percent gains is not worth it.
> Or just write faster JS using it's lower-level language features.
You are simply ignoring the history of these tools and why people have begun to reach for the languages in question.
I think there's a huge spectrum between "pokes around in source code occasionally" and "maintains critical infrastructure." Learning to read and understand the open source software that powers a lot of the toolchain is a big part of the development of a lot of engineers, and said toolchain mostly not being written in a language you're familiar with is a barrier to that (not an insurmountable barrier, but still a barrier, to be clear).
I agree with you that the juice is worth the squeeze here, but I don't think it's right to pretend that there's absolutely zero cost in terms of the learning pathway for the next generation of open source contributors and maintainers.
> I start wondering what we're supposed to do with all those precious milliseconds it just handed us back
Speed matters. WebAssembly is taking jobs from JavaScript precisely because it's faster.
The calculation engine for Google Sheets became twice as fast with the switch to WebAssembly: https://web.dev/case-studies/google-sheets-wasmgc
The Amazon Prime Video app became twice as fast with less variability in performance when they switched to WebAssembly: https://www.amazon.science/blog/how-prime-video-updates-its-...
Compiling to WebAssembly enables every language to run in the browser. Google used Java and Amazon used Rust.
It's late for me. I've commented plenty on this thread but I'll revisit this particular comment tomorrow with my assessment of these two blog posts.
Just skimming the Google Sheets wasm one, looks like pure technical debt chaos so I'm already plenty suspicious.
Suffice it to say that all these pseudo-technical but actually marketing blog posts should be taken with a truckload of salt. Apples-to-oranges unless proven otherwise.
Heterodox argument: a big part of the reason JavaScript has been so successful is that interpreted languages actually can be extremely fast—competitive with, and in many cases, faster than compiled languages, with V8 arguably the fastest interpreter in existence.
See this post by Mike Pall, author of LuaJIT (which is comparable to or faster than V8 performance-wise, despite being basically a single-developer project), which explains why: https://web.archive.org/web/20180603053407/http://article.gm.... Basically, it’s much easier to add high-quality runtime-trace-aware recompilation to a JIT interpreter, which a non-trace-aware compiled language will often not be able to beat.
Javascript only becomes fast because most of the CPU time is spent in native code rather than actual compiled or interpreted javascript.
The big diff between scripted and compiled languages is the former dispatches by name and the later dispatches by index. No surprises which is faster and which can be compiled in the faster native code.
JavaScript only becomes fast because it continually is making tradeoffs to decide what level of work to put into running any given bit of code. For a while this was three or four different targets on V8 (which I think maybe is down to just 2 now again).
Compiled vs interpretted is viewing this through the compiled language lens. I think the interesting part is that some runtimes are dynamically deciding how to run the code. The contrast versus those folks who decide and bake every single decision in ahead of time, where the only thing that happens is totally pre-scripted, is strong.
Stop Writing Dead Programs is such a lovely talk that hopefully can shake a couple of these sleepers awake, get us to see how absurd compiled code really is. And is quite fun and funnily said. https://news.ycombinator.com/item?id=33270235
It's mostly fast enough for most apps and that definitely helps. It's not as consistently fast as some languages that don't rely on an JIT. That matters more for tooling than your average website.
There's no need to even contextualize your argument, it is the reality. The whole point of spending such heroic efforts on a fast VM implementation is so that even faster peformance becomes a niche use cases.
But alas, many unskilled programmers use Javascript to make useful things - sometimes beyond their ability. Even a state-of-the-art virtual machine like V8 cannot hope to fix that class of peformance problems.
I will withhold my opinion on the "faster than compiled languages" part though. ;)
> Rust, Go and Zig are taking over increasingly large parts of the JavaScript toolchain...
Large parts of the JavaScript application space too.
Language consistency, ergonomics, standard library and performance matters, and JS has major warts here. I bet when these languages are 30+ years old like JS is, the software landscape isn't dominated nearly as much by JS.
These days I intentionally start all projects with as little JS as possible, opting for Go and HTMX instead. Removing the layers of JS inconsistency and build tools makes my and my agents lives better.
More thoughts on my JS-less stack here: https://housecat.com/blog/the-hugs-stack-hypermedia-unix-go-...
I love JavaScript and TypeScript, but like all interpreted languages, they really do have a hard time fitting into certain environments where you can't get an enormously complicated VM or can't get by with a much slower version of the language.
You can argue that high-throughput / low-latency dev tools are one of these environments. I'm not so sure. I think a lot of JS tooling is written in slow JS and gets fast with a rewrite to another language largely because of better structure.
Some environments really are tougher though. I started using WebAssembly a lot a year ago and the story for JavaScript and Python is just pretty terrible. And for JS, the things that make it hard to run in Wasm aren't universally good things to have in the language in the first place. The extreme dynamism of JS is something that so much JS tooling tries to limit.
But I think a lot of the language is quite good, and that there are possible variants of JS that are great even. I started one I hope will fit that one day (https://zena-lang.dev) and the first use cases for me are all tooling to try to make sure it's at least good for that: self-hosted compiler, formatter, regex engine, JSON parser, etc. It's quite a bit like JS/TS, but fixes a lot of things and runs forward with features that will take JS many years to get to, if it ever does (ie, pipelines, pattern matching).
I'm hoping that WebAssembly can slowly continue to gain a place on the web so JS can give up its near monopoly there. We'll see!
JavaScript is an abomination that never should've been left alive.
And yes, I ask myself the question: "Who in his right mind would run JavaScript on the server and even write business logic in it?" Lots of idiots in this world it seems. JavaScript is the reason our text editors need 16GB of RAM to run these days and a simple weather app 1GB.
In the ol' days assembly programmers probably could've written the weather app in a couple of KB (that's a MILLION times less memory people).
A programmer that can write a small efficient weather app in assembly can also easily do so in JS. Especially writing it in a c-style imperative way.
Modern Javascript has many low-level facilities and a great VM. You're conflating the low, average skill of the JS community to what the language is capable of.
People say this yet still are served some of the worse performing apps from supposedly the "best" corporations in the industry.
Low average skill unfortunately includes them as well. I say unfortunate because, like the build tool developers, they serve an outsized audience with their poorly written apps.
But, to be fair, large tech corporation have many other roadblocks to better software quality, well before you reach testing individual skill levels. Moral decay and office politics to say the least...
Exactly!
In todays of multi processors, its seems completely backward to adopt a runtime that is limited to a single thread.
JS has Web Workers in the browser and Worker Threads module in nodejs for parallelism. Most apps don't need to them to be fast, but many slow apps would be less slow if Workers were used more often.
Node is a single thread per process.
While other languages leverage sharing memory amongst multiple cpus, node requires a process per CPU.
From experience... It was really amazing to share TypeScript protocol definitions between the server and client. The client side would get all the autocomplete and intellisense and typed goodies, the server side used a compiler extension called typescript-is to validate the request bodies in one go, using the types as the source of truth. I loved not having to duplicate the validation logic in a separate way and manually maintain it to be consistent with the types. To this day I think that was the most natural way I've ever written a server-client pair and to this day nothing's compared, even supposedly cross-language things like protobuf. Ugh, TypeScript is so good at what it does. Sometimes I hate the lack of nominal types, but it really is up there with Rust as one of my favorite languages.
You can wait for slow programs that use a ton of excess memory, sure.
Or, you can have fast programs that use memory efficiently.
You just can’t do it in JavaScript. You can get really performant with JavaScript with careful engineering for sure… but there’s a big wall of diminishing returns even with all of the JIT compilation and tracing garbage collection.
Nothing stopping those same JS developers from learning some Zig or Rust or whatever. It might actually be easier to learn than memorizing all of JavaScript’s implicit coercion rules.
But if that’s too much to ask then stay in JS land and make sure there are JS devs to write tools for!
Instead of writing utilities, libraries, guidelines, build-time transformations, etc to help (or even) force JS devs to stay on the happy path, the industry "experts", large companies, and thought leaders chose to go the exact opposite route with React and adjacent tools. The results speak for themselves...
Even the V8 team gave up on documentation: https://v8.dev/blog
There is nothing stopping JS devs from working with byte buffers with near-zero overhead.
To all reasonable software developers out there, please don't listen to the prevailing groupthink. Javascript (the language) has many semantic problems but speed (from the VM) is one of it's best features! Like, you have to be writing some questionable code in very questionable styles/dialects to have a 10x (or more!) slowdown vs native.
Obviously - using native non-portable language/compiler features - you can reach some worthwhile speedup for certain workloads. But I have yet to see any of these "we rewrote our build system to Rust" type blog posts utilize any them.
It's always the apples-to-oranges marketing style drivel. Just because the some assertion comes from a very (very!) large company or well-known community member doesn't mean it is true.
In all my years of lurking this site, this port mortem is one of the few articles on this topic I trust:
https://zaplib.com/docs/blog_post_mortem.html
This argument can also be applied to the separate, startup time performance axis. Though there are more tradeoffs there.
I'm interested to know why tsc (pre-go) would give me near-instant LSP diagnostics on a large typescript project, and rust-analyzer takes ages to produce diagnostics in a small rust project.
If you need a desktop app go with rust, zig or swift depending on your platform
If you need a web app for CRUD, avoid frameworks, typescript, tailwind, nothing, just bare-naked node/html/css/js
If you need games in the browser or stuff where every byte needs optimization go with WASM
A programmer's life has become simpler
> just bare-naked node/html/css/js
Why include node here?
If you need to be pawned go with Node.
i enjoyed learning go. not that far from js
when i first started learning programming i thought it was super important to pick one language to be really good at and that could do everything but you dont have to pick just one.
i honestly think learning it opened my mind to thinking about programming differently and is probably better in the end
learning more languages is good actually. I still mostly write imperative code but having spent some time learning functional languages was very good for my brain.
I've worked with Vue, React, Qwik, and Svelte over the last 10 years. I have never felt the hate that HN spews out over JavaScript.
i don't mean to be a snark, but honestly I can't get the point of the long essay. What's the main idea?
tldr: interpreted languages are still slower than compiled languages, surprising no one
> Rust, Go and Zig are taking over increasingly large parts of the JavaScript toolchain...
Found it.