Hacker News Reader

Comment by anthonyshort | Hacker News Reader

I don’t understand the push to make userland component abstractions into native primitives. It’s perfectly fine for the browser to just implement low level primitives for frameworks like React to build on top of. It allows the framework to develop much faster.

Web components kind of miss a lot of the reasons why React has taken off. Encapsulation is one thing but it’s the declarative way of handling state, effects, and the DOM that make it powerful. Web components would need a framework on top of it to make it usable by developers building anything significant.

View on HN

They aren't trying to solve the same problems. The main problem they solve, that no other userland framework solves, is that I can now share a collection of my rich JavaScript components with every single web developer in the world regardless of the technology they build web apps with.

I can imagine a world where this would be useful, being able to drop something in as easy as a <select/> element seems fine.

It seems useful for UI primitives, like a dialog, tooltip etc, that the browsers might not implement as native elements themselves. But as a building block for creating entire applications? I don’t think so.

Which means they’re always going to need to hook into the framework that’s rendering them, either to connect state, managing effects, or for styling.

I just don’t see sharing a component regardless of framework as a major problem worth solving right now. Buy into a framework and community like React and you can just drop components into your app. It’s not framework-agnostic but it works.

I would rather have browser provide rich components then rely on third party.

That's one of the huge benefits of web components -- third parties can iterate quickly to find useful patterns, then browsers can just standardize the best/most popular ones. The old way of discussing ad nauseam in committee then implementing straight in the browser gave much less useful elements. (And it's much less straightforward to convert a popular React component into something browsers could implement natively.)

> hird parties can iterate quickly to find useful patterns, then browsers can just standardize the best/most popular ones.

I'm looking at hundreds of date pickers, carousels, autocompletes, customizable selects, data tables, modals (e.g. for media previews). They've been around for as long as there have been browsers (so, thrity years or so).

Can you point at any standardisation effort around any of these commonly repeated patterns?

input type="date" and input type="color" are only now barely making it into the browser.

Reply on news.ycombinator.com