Hacker News Reader

Comment by bluejekyll | Hacker News Reader

With WebAssembly, now you don’t have to. You do need a JS shim to load the WebAssembly, but after that, pick your favorite source language that can target wasm.

View on HN

This is fascinating. As someone who programs primarily in Python, I have been struggling with adapting to a JS-heavy environment in the past several years.

I have began utilizing Node + React for frontend use cases but find that my build pipelines become incredibly cluttered and esoteric rather quickly.

Am going to explore wasm solutions, thanks :)

I’ve long been of the opinion you must learn JS today regardless of what your preferred language might be - it’s more or less (thanks to browser vendors and the web) the closest thing we have to a “zero” dependency runtime on all platforms. Learning a WASM based workaround to avoid learning JavaScript is not helping you or the products you might want to make. There is a huge ecosystem of existing JavaScript code to adopt or extend in the client out there too. Expecting or wishing the market to bend to your preferred language or style is an often great way to become a very unhappy developer.

As someone who programs primarily in Python, I have been struggling with adapting...

Would you mind providing some more details here? After Python packaging drove me batty for the last time, I wouldn't have described the switch to JS/CS/TS as "struggling"?

With respect to "build pipelines", you don't have to use grunt, gulp, etc. It's totally fine to have regular bash commands in npm scripts.

Hm... I think the biggest issues I experience have more to do with frontend design rather than composing JS itself.

I love using it as a "one-off" scripting language which usually involves interaction with some sort of an existing codebase.

I'm also fairly confident in doing any sort of backend-centric work, in which existing components most likely exist as well (or at the very least something POC-esque to iterate on or scaffold from).

With that said I believe most of these difficulties stem from most of my experience being backend-centric roles, so when it comes time to implement a frontend from scratch, I become unsure of where to start. The sheer amount of technologies available and vast amount of flexibility is also very intimidating.

However I've been slowly but surely immersing myself in the stack(s), and have found that much like my experience with CSS the best way to learn is to simply do it-- making mistakes and learning from them in my exploratory learning.

I do appreciate what you said about build pipelines and feel much better about my current projects :-)

quick edit: I think my lack of experience in functional programming is another factor here, however Typescript's object orientated style is very attractive.

WebAssembly holds a lot of promise for sure, but the interface with the outside world still has to go through JS for now for anything non-trivial. I really do hope that this will change soon.

For web programming, WebAssembly is still not enough.

It solves the problem of how to run your non-web based code in a browser, but until it can interoperate with the DOM, JS will continue to be mandatory.

I think you mean efficiently interoperate with the DOM. There are plenty of abstractions (at least that I’ve worked with in Rust) that allow for dynamic management of the DOM.

For example, see the wasm-bindgen and the generated web-APIs in web-sys: https://docs.rs/web-sys/0.3.49/web_sys/

A nice framework that takes advantage of this is yew: https://yew.rs/docs/en/

Reply on news.ycombinator.com