RE: Throttling requestAnimationFrame based on viewport visibility

Sami,
Just checking back to see what was found from this investigation and follow-ups with Mozilla.

Is there anything that we should be improving in the specs for Page Visibility or the HTML5 reference for requestAnimationFrame related to this that will assist implementors?

-Todd

From: Ilya Grigorik [mailto:igrigorik@google.com]
Sent: Tuesday, August 11, 2015 4:49 PM
To: Sami Kyostila <skyostil@google.com>
Cc: public-web-perf <public-web-perf@w3.org>
Subject: Re: Throttling requestAnimationFrame based on viewport visibility

Hi Sami.

On Tue, Aug 4, 2015 at 7:15 AM, Sami Kyostila <skyostil@google.com<mailto:skyostil@google.com>> wrote:
The obvious compatibility risk is breaking pages that rely on an
out-of-view rAF to function. We're trying to avoid that by only
throttling cross-origin iframes -- with the hope that since there's no
synchronous way to observe rAF callbacks in those frames, sites are
less likely to rely on their timing. Does that sound reasonable?

I think the best answer to this would be to watch for developer feedback on Mozilla's implementation? :)

The motivation sounds reasonable, but it's hard to say what this might break. Also, as you pointed out in [1] singling out raF is a bit arbitrary.. curious to see what you'll learn from your experiments!

ig

[1] https://groups.google.com/a/chromium.org/d/msg/blink-dev/_SRHebxivJs/5zMt9tLdCQAJ