15:58:14 <RRSAgent> RRSAgent has joined #htmlt
15:58:14 <RRSAgent> logging to http://www.w3.org/2013/02/26-htmlt-irc
15:59:44 <krisk> Helo
16:00:21 <MikeSmith> hey
16:00:38 <krisk> we can wait a few more minutes to see if others attend
16:00:53 <krisk> Here is the agenda -> http://lists.w3.org/Archives/Public/public-html-testsuite/2013Feb/0021.html
16:04:16 <tobie> tobie has joined #htmlt
16:04:41 <krisk> Let's get going
16:05:31 <krisk> With the new 'folder' structure and robin
16:05:53 <tobie> Have to send my regret, unfortunately.
16:05:55 <krisk> ..robin's coverage report http://w3c-test.org/html-testsuite/master/tools/coverage/
16:06:04 <tobie> s/regret/regrets/
16:06:27 <krisk> I think it would be good to start to identify areas that we would like contributions from to the HTML test suite
16:07:03 <krisk> For example a Test The Web Forward event is comming up (Seattle) that this would be helpful to 'direct' contributors
16:07:38 <tobie> Is this meeting being minuted?
16:07:56 <krisk> yes ( note the RRS Agent)
16:08:08 <tobie> ty.
16:09:21 <krisk> thoughts?
16:11:32 <jgraham> Most of the areas where we really need coverage aren't the things that people are likely to pick at TTWF, I think
16:13:32 <mdyck> at TTWF, how do people pick? do they get any direction?
16:14:02 <jgraham> Things like "Browsing the web", embed, iframe seamless, and so on. Not really sexy features.
16:14:13 <krisk> It's normally pretty high level - test HTML5, CSS Border-Radius, etc..e.g. at the 'spec' level
16:14:35 <jgraham> mdyck: Based on my experience in Paris they do get some, but people naturally gravitate towards "new, shiny"
16:14:46 <jgraham> So WebRTC was very over-represented
16:14:54 <jgraham> and HTML under-represented
16:14:58 <krisk> yes, I agree that is the case...
16:15:11 <jgraham> CSS was pretty popular too
16:16:32 <krisk> Well I'd still like the this task force to start to enumerate areas of the spec we believe lack tests and have interop issues.
16:16:50 <jgraham> darobin has been doing some work in that area
16:17:05 <krisk> Even if it's at a 'section' level
16:18:21 <krisk> One way would be to add this to Robin's coverage report
16:19:52 <krisk> For example take section 6.1.2 enabling-and-disabling-scripting
16:21:31 <krisk> I believe IE, Chrome, Opera, Firefox, Safari don't have interop issues with this part of the spec
16:22:14 <krisk> ..and we do have sandbox tests, though they are located in another section.
16:23:15 <krisk> It seems good to have this reflected in the 'coverage' report, but today it shows up as a section with no tests
16:23:22 <Ms2ger> Ms2ger has joined #HTMLT
16:24:04 <jgraham> #enabling-and-disabling scripting is just definitions; there aren't any conformance requirements
16:24:13 <jgraham> So that seems like a bad example
16:24:41 <krisk> Don't focus on the specific example, per se
16:25:11 <krisk> Though I would add that if a section has no normative requirements we should also call this out in the 'coverage' report
16:25:44 <jgraham> Sure, a section without requirements won't have tests
16:25:55 <krisk> darobin are you following?
16:26:56 <mdyck> krisk: the coverage report already does kind of call it out: zeroes in the 2119, Algos, IDL columns
16:28:18 <Ms2ger> RRSAgent, pointer?
16:28:18 <RRSAgent> See http://www.w3.org/2013/02/26-htmlt-irc#T16-28-18
16:28:18 <mdyck> are there sections with those zeroes that *do* have normative requirements?
16:28:29 <Ms2ger> RRSAgent, make logs public
16:31:21 <krisk> the-draggable-attribute
16:31:32 <krisk> http://www.w3.org/html/wg/drafts/html/master/single-page.html#the-draggable-attribute
16:32:06 <Ms2ger> Has 7 2119 keywords, according to the report
16:34:00 <krisk> Ms2Ger did you read the IRC?
16:34:24 <Ms2ger> Sounds like I missed something? :)
16:35:51 <krisk> sounds like you did not
16:36:41 <mdyck> 6 "must", 1 "should"
16:37:15 <krisk> My point is that I'm sure some areas of the spec we may not have tests for but seems to be interoperable already
16:38:01 <Ms2ger> I don't think so
16:38:28 <Ms2ger> If you write sufficiently thorough tests, you're extremely likely to find differences
16:39:04 <Ms2ger> Whether those are important differences... Maybe, maybe not
16:39:30 <krisk> that is the key point - some interop differeneces are super important and others not as much
16:41:08 <krisk> If someone wants to go write tests I think it's good to at least offer a list of items that are known to be important and have interop issues
16:42:31 <krisk> For example I think having an appcache test would be more valuable than a test for 'the-datetime-element'
16:43:24 <Ms2ger> Or pretty much everything under editing
16:46:10 <krisk> mdcyk does that make sense?
16:50:44 <mdyck> sure. who gets to decide (or suggest) which testing areas are more valuable?
16:51:15 <Ms2ger> krisk? ;)
16:51:31 <krisk> in theory anyone of the participants in the testing effort
16:51:44 <krisk> ..and then people can agree or not
16:52:41 <krisk> fine to also ask/discuss with the whole HTML WG
16:53:44 <krisk> any disagreements?
16:56:33 <krisk> sounds like no objections
16:57:01 <krisk> it's getting close to the end of the meeting shall we adjorn?
16:57:39 <krisk> Note here is a good section we should priortize to have tests for 'the-input-byte-stream'
16:59:18 <mdyck> Is "valuable" distinct from "priority"? E.g., can you say "this is a valuable testing area, but it now has good testing coverage, so more tests there are not a priority"
17:00:10 <mdyck> Or is priority (as used in that sentence) the only useful thing to convey?
17:00:45 <krisk> all of which...
17:00:58 <krisk> for a given section we could mark it as...
17:01:09 <krisk> A) no requirements - done
17:01:58 <krisk> B) has conformance requirements, no tests, known interop issues (for example the-input-byte-stream)
17:02:50 <krisk> C) has conformance requirements, has tests
17:03:25 <krisk> Such that it would a high 'priority' to have tests first for
17:03:38 <krisk> ...'B'
17:04:07 <krisk> it's nice to have tests for 'C' but not if they are not finding interop issues
17:04:19 <krisk> make sense?
17:06:12 <mdyck> well, A|B|C aren't exhaustive, leaves out: has conf reqs, no tests, no known interop issues
17:07:15 <krisk> then say add 'D'
17:07:36 <mdyck> also, does "known interop issues" take the place of "valuable" discussed before?
17:07:37 <krisk> D) has conf reqs, no tests, no known interop issues
17:09:11 <krisk> agree?
17:09:52 <mdyck> The coverage report already (tries to) convey whether section has conf reqs and tests, so wrt your markings, what it needs is to convey level of known interop issues
17:11:03 <krisk> yes
17:12:22 <mdyck> i wonder if coverage report could pull data from caniuse
17:12:34 <krisk> no
17:21:33 <krisk> we are far past the scheduled time for the meeting, need to run
17:21:49 <krisk> let's end the meeting
17:22:00 <krisk> RRSAgent, make logs public
17:22:06 <krisk> rrsagent, generate minutes
17:22:06 <RRSAgent> I have made the request to generate http://www.w3.org/2013/02/26-htmlt-minutes.html krisk
18:55:42 <gitbot> gitbot has joined #htmlt
18:55:42 <gitbot> [13html-testsuite] 15rhauck opened pull request #33: Submission/rhauck (06master...06submission/rhauck)  02https://github.com/w3c/html-testsuite/pull/33
18:55:42 <gitbot> gitbot has left #HTMLT
19:57:55 <darobin> darobin has joined #htmlt