- From: Nat Duca <nduca@chromium.org>
- Date: Fri, 13 Mar 2015 11:29:13 +0900
- To: Boris Zbarsky <bzbarsky@mit.edu>
- Cc: "public-web-perf@w3.org" <public-web-perf@w3.org>
Received on Friday, 13 March 2015 02:30:13 UTC
In line with what Boris says, it is very hard for us to know when a window is covered by another window. Not every OS is consistent on telling windows about occlusion --- it is computationally expensive to figure out occlusion. We can know in some extreme cases: minimizing, for instance. But, there are a lot of corner cases: when a window goes to a hidden workspace, for instance, some OS/window managers tell us, but some don't. The page visibility spec is, I recall, crafted carefully, to allow false positives for "visible" but not false negatives. That is, "page is visible" means "the page may be visible to the user, but user agent doesn't have to be 100% sure." But, "page is not visible" means "the page is definitely not visible to the user." This is the very important one for us to get right.
Received on Friday, 13 March 2015 02:30:13 UTC