25428 2014-04-23 13:52:59 +0000 hasFeature removal was premature 2015-08-03 13:02:13 +0000 1 1 1 Unclassified WebAppsWG DOM unspecified PC All RESOLVED INVALID https://bugzilla.mozilla.org/show_bug.cgi?id=984778 P2 normal --- 1 philipj annevk bugs crimsteam ehsan mike www-dom public-webapps-bugzilla oldest_to_newest 104265 0 philipj 2014-04-23 13:52:59 +0000 hasFeature was made to always return true on my request. It looks like I underestimated just how much the use counter data can change when it reaches the stable channel: http://www.chromestatus.com/metrics/feature/timeline/popularity/230 http://www.chromestatus.com/metrics/feature/timeline/popularity/231 Removing hasFeature entirely is hopeless, and making it always return true is also right above the highest threshold (0.06%) ever cited for Blink. It would be very interesting to learn what the features that return false are, but there's no way of collecting strings with the use counter system... Sadly, I think the path of least resistance is to just go back to special-casing SVG. 104739 1 philipj 2014-04-29 22:11:27 +0000 Huh, something strange has happened in http://www.chromestatus.com/metrics/feature/timeline/popularity/231 The counts dropped from above 0.06% to around 0.004% between April 24 and 26. I have no idea how that could happen, will consult blink-dev. 104794 2 philipj 2014-04-30 09:30:39 +0000 Here's the blink-dev thread: https://groups.google.com/a/chromium.org/d/msg/blink-dev/yCUjWJAUd9g/89KL8HjApecJ Since it's not clear what's going on, I suggest waiting for a few weeks or months before drawing any conclusions from the data. In the meanwhile, Anne, do you think Mozilla would be willing change hasFeature to always return true if the usage turns out to be really low? 104800 3 bugs 2014-04-30 09:51:37 +0000 Would need to add some telemetry. Since the special cases are for svg, even low usage could mean breaking quite large portion of svg content. So better to have telemetry for svg part only and compare true and false return value counts. 109532 4 philipj 2014-07-29 22:52:40 +0000 Chrome's use counter has stayed around 0.004% until now. In Chrome, SVG is used on around 10% of page loads, so even relative to that it's look good: http://www.chromestatus.com/metrics/feature/timeline/popularity/250 Any luck with telemetry in Firefox? 122402 5 philipj 2015-08-03 13:02:13 +0000 hasFeature() always returns true now, https://codereview.chromium.org/1108293002 was in Chromium M44, and Chrome 44 was recently released. I'll close this and revisit only if there are regressions reported.