cross-posted from: https://programming.dev/post/1379446
Why I think Typescript is A Good Thing, But mainly a complaint about the number of programming languages. Other than objects (which were not invented), I don’t think any of them are a marked improvement on PL/1.
JavaScript’s longevity is assured for one reason. Browsers only support JavaScript.
I do agree that JavaScript will be with us for quite a while still, but that statement has a caveat: WebAssembly
You can use a backend language, for example Rust has good tooling here, and compile it to WebAssembly, which runs in browsers.
That way, your frontend stack can be HTML+CSS+Rust.This also has various advantages:
- Better performance than JS.
- Can share the communication model between backend and frontend.
- Don’t have to write JS (or TS, which still carries over quite a few oddities from JS).
No, you can’t replace JS with Wasm entirely. Wasm can’t access the DOM https://developer.mozilla.org/en-US/docs/WebAssembly/Concepts
It’s not a perfect solution yet, but their comment is mostly accurate. wasm-bindgen works well enough to create whatever else is missing. In my opinion, it can feel annoying to use when doing stuff like managing callbacks with rust, but, there are web frontend frameworks that are progressing.
I’ve used Dioxus. If the application isn’t too complicated, it’s not a bad experience. It’s still in development though. There’s also Yew, which has a similar react like design that Dioxus uses, but I preferred Dioxus. Lemmy also actually has a frontend client that uses Leptos, which is another react like framework in rust.
If someone was using egui, I don’t think they would have issues with their user interface on the web. (but be more constrained to the limitations of it)
—-
They were also only correcting the author who stated browsers support only JS. I mean, I’m not sure I would recommend rust to a new developer, but, I think they’re correct that browsers support more than JS. (Browsers are just not supporting WebAssembly to the level of JS yet)
I’m using Leptos myself and my understanding is that it’s decisively different from React in that it does not maintain a Virtual DOM. In that sense, it’s more reactive than React.
Well, unless you meant something different. I don’t know terribly much about React. 🙃
As I understand things, yes, WASM has to interop with JS to handle the DOM access. But I have implemented a small UI with DOM access and had to write no JS myself, because frameworks can generate that interop code for you.