3 ms·
I think the problem is deeper in the standard. Funding those who run after a train that is designed to be expensive to catch up with is wasted effort. We need
by wruza 2y ago
I think the problem is deeper in the standard. Funding those who run after a train that is designed to be expensive to catch up with is wasted effort.
We need Web-next which would be clearly defined in proper primitives and features that don’t suck from the beginning and need no further extension. And a reference implementation of it in a clean way, even if not very performant. E.g. fine if it takes 50MB of RAM just to start and show a welcome tab. Implementors will optimize it later.
- high_na_euv 2y agoWhy you need reference impl?
- wizzwizz4 2y agoThere are things you figure out when writing / playing with a reference implementation that aren't obvious from just the spec. For example: the W3C thought `box-sizing: content-box;` was a good idea, but Internet Explorer 3 implemented `box-sizing: border-box;` semantics which are, just, obviously better. (IE6 changed to context-box, but the W3C were introducing border-box in CSS3, which was then available in IE8.)
- amelius 2y agoYes, and in a purely functional language like Haskell.