- From: François REMY <francois.remy.dev@outlook.com>
- Date: Thu, 10 Jan 2013 00:49:00 +0100
- To: Paul Bakaus <pbakaus@zynga.com>, CSS WG <www-style@w3.org>
- CC: Tobie Langel <tobie@fb.com>, Tab Atkins <jackalmage@gmail.com>
> Hi François, > > Love your proposal, very intriguing. Not sure if I should be concerned or > excited that it seems that you (well, more-so Tab ;) would like to create > a "high performance CSS alternative" within CSS. But hey, if it helps > performance and is relatively straight forward to implement for the web > developers, all for it. This is actually a point where Tab and I do not completely agree. I do believe that a judicious use of the current CSS layout engines and of local viewports will be sufficient to create amazing (layout/scrolling) performances. I understand I'll have to prove this someway (ie: prove that you can get the benefits of a viewport-canvas using viewports and any other layout engine). This is a nice challenge, and I have the big picture in mind already, but proving this would require making a comparative experiment and we don't have any 'pure canvas' implementation to test against. I guess I would need to make a benchmark to compare against current performance, but I'm not sure what to test exactly and how to enable 'off-screen elements non-rendering' optimizations via scripting. > Looking forward to hear Tab's finding on whether such optimizations help a > lot or just a little. Me too ;-)
Received on Wednesday, 9 January 2013 23:49:28 UTC