"will-change" is being a mess right now. Still wary of it as a CSS property. Feels like a band-aid. buff.ly/2dI3qIm
3
3
8
These problems are caused by Chrome's implementation choices, not by the property itself. Please complain to the Chrome team.
2
1
They've had enough complaints. I've been questioning will-change's legitimacy as a CSS property for awhile now.
1
1
Everybody agrees that the property shouldn't be necessary in a perfect world. That's not a contested point.
1
1
Nevertheless, until rasterization perf becomes a non-issue, will-change allows real perf gains on real websites today.
1
1
When will it become a non-issue? What conditions do we need to meet?
1
Hard to say. It requires browser engines to be just as fast at re-rendering everything as they are at recompositing layers.
1
What's keeping that from happening? Are there hard constraints we might not be able to overcome?
2
Replying to @rachelnabors @markusstange
Maybe graphics cards designed for Web rather than games, plus Web tech designed around graphics card limits?

Oct 6, 2016 · 6:01 PM UTC

1
Replying to @davidbaron @markusstange
How do we design around graphics cards limits? Something I keep running into: we have no way of knowing capacity.
1
1
I meant the @csswg designing around those limits, not webdevs.
1