- From: Robin Berjon <robin@w3.org>
- Date: Thu, 18 Sep 2014 12:26:13 +0200
- To: Adrian Roselli <Roselli@algonquinstudios.com>, Jirka Kosek <jirka@kosek.cz>, John Foliot <john@foliot.ca>, 'Daniel Glazman' <daniel.glazman@disruptive-innovations.com>, "chaals@yandex-team.ru" <chaals@yandex-team.ru>, 'Philippe Le Hegaret' <plh@w3.org>
- CC: "public-html@w3.org" <public-html@w3.org>
On 17/09/2014 16:00 , Adrian Roselli wrote: >> From: Robin Berjon [mailto:robin@w3.org] Hi Adrian, > > Hi, Robin! Hi! >> It is out of the scope of this WG in the sense that this WG isn't >> responsible for the whole Web Platform Tests project. But this WG >> is, at least theoretically, a stakeholder in that project and has >> an obligation to produce tests for it. So it's a question worth >> asking. > > Fair enough, but I don't think the tests that are written here to > verify that a feature has been implemented (ideally correctly) are at > all what developers come to find when looking for browser support. > Partly because the tests aren't ongoing and don't track changes in > support over time (bugs, tweaks, etc) which seems hard to do in the > new constant release cycle some have adopted. The tests in WPT are ongoing and continuously updated. I agree that most developers won't care much about the results at the granularity level of a single test (that will generally describe one alternative in one case of one type of parameter given to one method of one feature) but since we're planning on providing the raw data people can infer useful information by aggregation. For a given feature you can see that Zinc might have 82% support, Tubes Visitor 84%, etc. Also, using SauceLabs we can get results over time since we can run the tests no pretty old releases. >> The goal isn't to replace CanIUse, but rather to publish >> high-quality, fine- grained test results for our specs. CIU is nice >> but it is in places woefully incomplete. If CIU wants to take our >> data and slap a nice interface in front of them, that's great. > > Agreed on all points, but don't discount that CIU is a de facto > resource I see referenced again and again for tracking browser > support, not W3C. > > So all that being said, does W3C want to become that leader in > tracking ongoing browser support for all features defined across all > its specs? We do plan to provide comprehensive test result data across all of our specs for all browsers we can get. That doesn't mean necessarily competing with CIU, it just means (at least) providing the data. If CIU wants to use this data to improve its accuracy and completeness, everyone wins. -- Robin Berjon - http://berjon.com/ - @robinberjon
Received on Thursday, 18 September 2014 10:26:38 UTC