On Tue, Jan 23, 2024 at 2:31 AM Domenic Denicola <dome...@chromium.org> wrote:
> > > On Tue, Jan 23, 2024 at 7:55 AM Thomas Guilbert <tguilb...@chromium.org> > wrote: > >> Good point about the most used APIs being the boolean properties! The >> APIs are now only aliases for the standard non-prefixed fullscreen APIs (see >> this code for the current implementation >> <https://source.chromium.org/chromium/chromium/src/+/main:third_party/blink/renderer/core/html/media/html_video_element.cc;l=418;drc=25705537e65f031d28c1a531046b11914055b7e6>), >> and therefore aren't much of a burden to maintain. >> >> I opened a WebKit standards position here: >> https://github.com/WebKit/standards-positions/issues/306 >> >> I unfortunately do not have access to edit the listed ChromeStatus entry, >> and the current owner no longer works on Chromium. Should I create a new >> feature (titled "Deprecation of HTMLVideoElement-specific Prefixed >> Fullscreen API")? I think the current ChromeStatus entry also covers this >> API >> <https://source.chromium.org/chromium/chromium/src/+/main:third_party/blink/renderer/core/fullscreen/element_fullscreen.idl;l=19;drc=047c7dc4ee1ce908d7fea38ca063fa2f80f92c77> >> which I am not trying to deprecate in this Intent. >> > > Yeah, I think a new feature is a good idea. The old feature seems to be > for the *addition* of the prefixed fullscreen API properties back in > Chrome 15, whereas a *deprecation/removal* has a different set of fields, > if I understand correctly. > > >> >> What's a reasonable sample size of HTTP Archive sites to audit? Should >> this be a complement/precursor to the proposed Deprecation Trial, or would >> this sampling be enough? >> > > I recall +Philip Jägenstedt <foo...@chromium.org> having done some > computations in the past based on what would actually be statistically > significant. But, I couldn't find them in this documentation > <https://www.chromium.org/blink/platform-predictability/compat-tools/> (or > the linked document > <https://docs.google.com/document/d/1cpjWFoXBiuFYI4zb9I7wHs7uYZ0ntbOgLwH-mgqXdEM/edit#heading=h.1m1gg72jnnrt>). > So, I'll just state that I would be happy with 20 sites. > You're probably thinking of these threads: https://groups.google.com/a/chromium.org/g/blink-dev/c/gIyvMw0n2qw/m/CC7RlhguAgAJ https://groups.google.com/a/chromium.org/g/blink-dev/c/V7q43bgutbo/m/GpJogpw4AgAJ Checking 20 sites at random and finding no breakage gives us 95% confidence that the true occurence of broken sites is no more than ~17%. Checking 40 reduces that to ~9%. (Based on https://sample-size.net/confidence-interval-proportion/.) If the set we're sampling from is sites that hit a use counter, then it says something about the real risk of breaking that code path. Checking 20 like Thomas did seems enough to me, we're starting from a low risk and just want to know if breakage is common or not. It seems to not be common even on sites that use the APIs. -- You received this message because you are subscribed to the Google Groups "blink-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscr...@chromium.org. To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAARdPYd7X6deDxQPywsiGJ3W%3DQPhTuPL8jQaB860SXZB0hxYVQ%40mail.gmail.com.
