Hacker News new | ask | show | jobs
by semiquaver 1025 days ago
Crossing my fingers that the proposal for allowing (browser-ignored) type annotations in javascript progresses: https://tc39.es/proposal-type-annotations/

Between that, HTTP2/3 and ES modules many of the downsides for building apps with no compile step are almost completely mitigated.
2 comments
nwienert 1025 days ago
Does http2 solve waterfall imports? Also wouldn't tree shaking still be unsolved? What about minification?

IMO build step woes are really overstated, my build step has linting, type checking, optimization, bundle analysis size diff checks, enables JSX, and is basically necessary for development anyway as you get hot reloading, compiler-added features for better debugging, etc.

link
mixedCase 1024 days ago
>Does http2 solve waterfall imports?

It used to, but browsers ruined it because they didn't want to help improve caching support

link
tomjen3 1024 days ago
HTTP2 allows the server to push files to the browser.

If you don't have tons of code, tree shaking and minification is not that important. If you depend on a couple libraries (Jquery, for example) those can be minified and just refered to directly in the source.

link
spiffytech 1024 days ago
Chrome removed support for HTTP/2 Server Push last year.

As I remember, it never saw broad adoption because it was technically challenging to integrate with, and it tended to push a lot of resources clients already had, so it wasn't a big win for efficiency anyway.

link
semiquaver 1024 days ago
HTTP early hints still exists and can achieve similar things as server push without wasting bandwidth.
link
ryncewynd 1025 days ago
Care to share your build set up? I'd very much appreciate it

I haven't done js since jquery days, but keen to get into modern js development

link
nwienert 1025 days ago
I maintain Tamagui, so the tamagui starter kit basically is it, try `npm create tamagui`
link
semiquaver 1024 days ago
Waterfall imports are solved by rel=modulepreload.
link
benatkin 1025 days ago
Wouldn't that conflict with possible future optional static typing? It's only Stage 1. I'm hoping it doesn't make it in its current form. Some good discussion in the issues, though: https://github.com/tc39/proposal-type-annotations/issues
link
Someone1234 1025 days ago
Can you explain why the two proposals would "conflict?" Seems like they're proposing the same thing to me.
link
benatkin 1025 days ago
It would be hard to separate ignored ones from ones that would use a possible future optional static typing system. They would possibly have to use a different, uglier syntax.
link
Someone1234 1024 days ago
I'm not sure I follow. Both proposals are suggesting that browser's engines ignore the syntax completely to allow some other, external, process to check them.

They use slightly different wording but are describing the same basic idea: This syntax will be ignored entirely and then optional/other process can use this area for whatever.

link
benatkin 1024 days ago
There are other languages that have optional static typing that don't treat types as comments. JavaScript could evolve to add support for this.
link