Meeting minutes
This meeting
Nigel: Lots to cover today - IMSC 1.2 feature requests, IMSC 1.2 CR CfC, TTML2 CR Tests,
… and DST switch.
… Is there any other business to cover, or particular points to cover not listed on the agenda?
group: [none]
IMSC 1.2 New feature requests and CR CfC
Nigel: We have recently had 3 new feature requests for IMSC 1.2.
… We need to decide if we go for CR based on the IMSC 1.2 FPWD or do editorial work
… to support some of these feature requests.
Cyril: Timeline of Rec - assuming most aggressive schedule, is Rec in 4-5 months
… reasonable?
<atsushi> https://w3c.github.io/spec-releases/milestones/?fpwd=2019-11-28
Nigel: 28 days for PR from CR, assuming we have no tests to complete, then a few weeks
… to REC. The hold up might be the call for exclusions period.
Atsushi: The Call for Exclusion period is 60 days. For IMSC 1.2 this would be completed on April 26.
Nigel: In that case the period would be 2 months before Rec.
Cyril: Assuming all transitions would go well.
… I think the CfC still has a week to run, then the transitions have to happen.
… From a Netflix perspective if IMSC 1.2 is going to be published in more than 3 months (roughly)
… then we will not insist on having the feature in IMSC 1.2 because we will not be able to change our plans anyway.
Nigel: Confused - do you mean less than 3 months?
Cyril: let's say we add fontShear to IMSC 1.2 today. If Rec publication would be more than 3
… months from now it wouldn't fit Netflix timeline so we would not insist on that.
… We still want the feature but we would wait for the next version.
… Maybe I was a little over-optimistic before.
Nigel: For me that is a clear steer that we should proceed with IMSC 1.2 as is, and
… soon open up for new requirements for IMSC 1.3 including the three recently raised,
… which simplifies things a lot. Any views from others in the group?
group: [silence]
Nigel: I think silence means everyone agrees that is the right answer.
Pierre: Yes.
Nigel: Thanks for reminding us of the CfC Cyril. That has 1 week to run, and currently
… no objections.
Call for Consensus to publish IMSC 1.2 CR
Nigel: We now have the choice of iterating through the three new issues to register that
… we will consider them for the next version.
Pierre: I don't think we should rush ahead with those - there may be impacts on TTML2
… and CSS, so we cannot make a judgement call now.
… There may be some that we decided not to do in the past because they are too hard.
… We need to come prepared to discuss them with plenty of time.
Nigel: OK, this is the sort of thing where a face to face meeting (if we can have those)
… would be quite useful.
Nigel: I will add a note offline to the new issues saying we mentioned them in today's
… call and will consider them for future versions of IMSC 1.2, etc.
Pierre: Thanks.
Nigel: The status of the CfC at the moment is that it still stands.
TTML2 2nd Edition CR Publication - Tests
Nigel: Thanks to everyone, chiefly Glenn, who has worked on the ttml2-tests issues.
… There are some where we need to agree we don't need tests, and others where I've
… been insistent that we do need tests.
Glenn: I can accommodate the request for those tests.
Nigel: That makes me happy.
Glenn: I can make one of the issues marked as not testable as testable.
… I've gone through and created issues for them and will now make a pass at
… creating tests - that's my plan for this coming week, and I will try to deal with
… the reopened issue 200.
… If I get through those then I will also make a stab at getting one set of implementations
… done on TTT/TTV/TTPE so we'll see where that stands this time next week.
Nigel: I wonder if we can seek group agreement for things we don't need to test right now.
Glenn: Unless someone disagrees, then there are no objections.
Nigel: I agree, let's highlight the issues are open and ask people to review them.
Cyril: Can we recap on the status?
Glenn: No new PRs. I created issues for all the TTML2 PRs except those I marked as
… untestable.
Cyril: How many were untestable?
Glenn: I think maybe 1..2..3 - ttml2#1099, ttml2#1109 and ttml2#1096,
… and that's it at the moment.
Cyril: I agree they are not testable on those three.
Nigel: I think I did as well.
Glenn: Overall I created 7 new issues, and I will proceed with creating PRs in the ttml2-tests
… repository specifying new tests.
… In most of those issues I created a comment that suggested a possible test approach.
Nigel: For at least one where there wasn't such a comment I added one.
… I'd ask everyone to review the 2nd Ed changes and verify that they're happy with the
… proposed test approach.
Glenn: Thanks for that. I see you also raised an issue on TTML2, and I found another editorial one, so we have
… a couple of changes there.
Tests in general
Cyril: Question about tests in general. When I was reviewing the tests, especially all the
… IMSC 1.1 tests I discovered we have lots of places where there are tests.
… I wonder if after this effort of going to CR we could create a single repo maybe with
… git submodules, and then annotating which tests should pass in which version of which
… spec would be really helpful to implementers.
Glenn: Those tests are not for implementers so I would say a big N O.
… The sole purpose of the tests is to pass CR.
Cyril: Still a single place for the tests would be really useful.
Glenn: I agree but tests for implementers is a completely different effort and a significant
… effort too. FYI the structure of the tests there now are mostly bound to at least some
… implementations used during the verification process to get to Rec and it would let's say
… break downstream systems for verifying them, e.g. TTT if we moved them.
… If we want to maintain TTV being able to perform verification on TTML1 and TTML2 it
… would break things. If we don't care about maintaining TTV then that's a different story.
Nigel: One path out of this would be to filter and copy existing tests into a new place,
… and consider the marketing value of doing that - for example we may want to put them
… into Web Platform Tests.
Glenn: We should think about why we are doing this - I think it is a worthy project to make
… interoperability tests but we would need to know who would back and fund it.
… The idea of web platform tests is a good one. That project if I recall was backed by a whole
… slew of organisations and funding from facebook and others if I recall correctly, to build
… the infrastructure to do that. It's a nice idea, but it doesn't sound right to me to take
… what we have now and restructure it.
AOB - DST switch
github: https://github.com/w3c/ttwg/issues/103
Nigel: Looking at the votes on the 3 options we have 4, 4 and 3.
Andreas: I signalled my presence but want to give a stronger weight to people
… who regularly join and are more involved. I see for example that Cyril has a clear
… preference for option 1. For me option 2 is better but option 1 is not impossible and
… also I have to admit that I often have blocking issues, if that helps to make a decision.
Nigel: Thank you for that.
Cyril: I have a preference for option 1 but could live with option 2 if needed.
Nigel: What do you think Gary?
Gary: I would like to leave it eventually in my local time because of a block of meetings
… on Thursday mornings.
Nigel: I think it's important that Chairs can attend!
… I could make it either way.
Cyril: One thing to bear in mind is that we sometimes have a 2 hour session. Which is easier
… in that case?
Nigel: For me that makes no difference.
Gary: If we extend an hour after that might be easier.
Nigel: The hour before was a historical thing from when we used to begin at 1500 UTC.
… (actually 1100 Boston)
Atsushi: I have a clash an hour before so it would be easier if we make 2 hour meetings
… extend later.
Nigel: Shall we say option 2?
Gary: There don't seem to be any objections.
Nigel: There were no down-votes. So that's done.
… Thanks everyone, I will adjust the meetings to suit.
SUMMARY: Option 2 selected.
Nigel: This means that for all the remaining meetings in March the start time will be 1600 UTC.
Pierre: Can I suggest you send an Outlook invite for the next few meetings?
Nigel: Will do.
Pierre: Thanks.
Glenn: On 2nd April it will be one hour later?
Nigel: No, earlier, 1500.
Atsushi: I will check the TTWG website too.
Nigel: And we should reissue the joining instructions too.
… The webex schedule might need adjusting too - I will cover that offline.
Meeting close
Nigel: Thanks everyone, let's adjourn there, having completed our agenda.
… Meet same time UTC next week. [adjourns meeting]