WebAssembly

malmon

Seems pretty neat, though I'm yet to actually find a decent tutorial on how to use it. One does it with C, another does it with JS, and another does it with plain bytecode. I guess it's still in its experimental stages, so there isn't really any full concensus on how to use it properly?

Also, WebAssembly general.

KuroAku

Looks promising. I don't like it very much from a security perspective as it increases the attack surface, but from a engineering perspective is a big progress.

There's this other project (http://asmjs.org/) that compiles code to a subset of JS that is very efficient. I don't know how compares to wasm, probably wasm is faster as is already compiled instead of VM efficient JS. I bet some web experts had already run some benchmarks on it.

Steph

I really like the idea of wasm, because it means we can finally get away from JS on the web. With LLVM support I should be able to run Crystal on the web, with a ported stdlib, which is pretty cool.

malmon

RX14 wrote:
I really like the idea of wasm, because it means we can finally get away from JS on the web. With LLVM support I should be able to run Crystal on the web, with a ported stdlib, which is pretty cool.

This annoys me though: "WebAssembly is designed to be a complement to, not replacement of, JavaScript."

http://webassembly.org/docs/faq/

Steph

malmon wrote:
RX14 wrote:
I really like the idea of wasm, because it means we can finally get away from JS on the web. With LLVM support I should be able to run Crystal on the web, with a ported stdlib, which is pretty cool.

This annoys me though: "WebAssembly is designed to be a complement to, not replacement of, JavaScript."

http://webassembly.org/docs/faq/

If it's possible to have a language work on the web without compiling to js, and being able to use all the JS apis, then you can get away from JS on the web. WebAssembly probably won't be a replacement to JS in the context of the whole web, but it can be a replacement to JS in the context of a single website/team.