Meeting minutes
This meeting
Nigel: Today, we have a quick update on DAPT-REQs
… TPAC Planning
… Rechartering,
… We also have TT in low latency, and Behaviour with controls. Drop them, or is there something to discuss?
Gary: Can probably drop them
Nigel: OK
… Any other business, or points to make sure we cover?
No other business
DAPT REQs
Nigel: We now have published the draft WG note, at
Nigel: And I saw that Atsushi merged the pull request that means whenever
… we merge a pull request to the default branch, it will trigger republication automatically.
… So as agreed, any PR on that document is considered a Call for Consensus.
… Just like with IMSC-HRM.
… I've prepared a blog post about it which Cyril has kindly looked over.
… Thank you.
… This is just something Chairs can do, so after this call I'll give it another look over and then publish.
… It's basically a call to action to review and contribute.
Andreas: General question about the DAPT profile:
… Is there any relationship to IMSC?
… I understand it will be similar to IMSC, and written in the same style.
… Any intent to take IMSC as a reference for the features included or excluded?
Cyril: The intent is not necessarily to represent something that can be rendered directly [visually]
… but there's an option to be able to associate styles and rendering with content.
… Two links with IMSC. One is that the structural constraints in terms of timing,
… attribute restrictions etc will be based on the same thing.
… Then for visual rendering, we want to base those semantics on IMSC too.
… It should be possible, if the author of the script also provides rendering indications then
… it should be easy to produce IMSC subtitles.
Andreas: That makes sense, thanks.
Nigel: I was going to say something similar.
… We may get some comments about timebase, I'm not sure,
… but I think we start with the position of media time only.
… I think the next stage is to start drafting the specification.
Cyril: Yes, the action is for me.
Nigel: Any more on this topic?
No
TPAC Planning
Nigel: Gary, will you be able to fill in the form for the Chairs?
Gary: Yes
Nigel: [runs through list of questions]
… Agenda topics? At the least, DAPT.
Cyril: If we have a meeting, I can show up.
Nigel: Me too
… Any other likely attendees in person?
Pierre: If there's something interesting on the agenda, then covid permitting I will be there,
… completely driven by agenda for me.
Nigel: Thank you. Any more?
… In that case, for now, in-person is 3 for the form.
… Any constraints on days or times?
Pierre: What are the dates?
Gary: September 12-16
Pierre: IBC is concurrent, so for me,
Cyril: IBC is 9-12
Pierre: Right, so at the earliest, 14th allowing for travel, for anyone there.
… Preference is more towards the end of the week than the beginning.
Nigel: That's a useful data point - it may affect others too.
… Next question is about group overlap and joint meetings
… Maybe Media WG because of controls?
Gary: Yes
Cyril: Interested in both Media WG and TTWG so would like to avoid non-deliberate overlap,
… but also having meetings close in time/day would be useful
Gary: Same here
… Any other groups to avoid overlap with?
Nigel: Sounds like a no
Andreas: There's the MEIG - would not be good to overlap
Nigel: Good point
Gary: That's already likely to avoid overlap because Chris can't attend both at the same time
Nigel: One from me: I've suggested to the Audio Description Community Group that we should
… hold a meeting, the intent is to gather input and build momentum for DAPT amongst non-TTWG people.
… (and maybe non-W3C members too)
… From a time of day perspective, I think we need to support
… Vancouver, Europe and Japan, based on the responses so far.
Atsushi: I don't think there's a sweet spot for Vancouver and Japan
Gary: We may have to do it early in the day
Nigel: Any other events of interest?
… Technical plenary with high level presentations
nobody
Nigel: I quite like that, myself
… Demos
nobody
Nigel: Hackathon
nobody
Nigel: Workshop
Pierre: Just a thought - TTML and WebVTT are used by a large number of people.
… Which is awesome. Very few of them are in this group.
… One option, to increase in-person meeting value, is to ask the question beyond this group.
… Are there hot topics? We could try to have a workshop with users.
… Lots more work but more productive and fulfilling.
… Want to throw that out. If we want to try to answer bigger questions or make
… progress on larger issues, we should cast the net beyond this group.
Nigel: Really good point.
Gary: To add, a lot of those larger issues apply equally to TTML/IMSC and WebVTT
… even if the implementation details might be different.
Pierre: Totally agree, a lot of it is people coming to terms with Timed Text
… Maybe TPAC is an opportunity.
Nigel: OK let's tick the Workshop box and bear this in mind. I can see it may well be worth the effort.
… Next one is Developer Meetup in the evening including local community
nobody
Nigel: Training overview on W3C technologies
nobody
Nigel: Any others not mentioned?
None
Nigel: Thanks. Gary, have you got everything you need?
Gary: I think so. Still unsure about timing, but I think we don't have to be specific yet.
Nigel: True - also worth noting the locations of likely meetings.
Gary: They also ask about meeting outside regular meeting hours, which I think we would have to do.
Pierre: Atsushi, how likely are you to be able to travel?
Atsushi: For now, quite unlikely due to requirements for re-entering Japan, unless they are relaxed.
… Please set the meeting time for convenience at the venue - I should adjust in any case.
Pierre: Let's say that we have a workshop. I think we might actually want to avoid having a group meeting at all.
… We can have those whenever we want.
… If we are going to miss important participants because of the time, maybe we can
… not have a TTWG meeting but instead have a workshop that benefits from in-person discussions.
… It would be silly to go to Vancouver to have a meeting we can do over the phone.
Nigel: I broadly agree, but note that there is an intangible value to making some decisions
… in person around the table, thinking back over previous meetings.
… Even if they could have been done over the phone.
… Anything else about TPAC for now?
Rechartering status update
Nigel: News!
… Two meetings, one with each formal objector.
… The discussion with Google resulted in a pull request that would resolve their objection.
Proposed edit to resolve Google FO
Nigel: Please review.
… Thanks Pierre and Gary for already approving.
… This PR changes the "For example" sentence in §3.1 Success Criteria
… which, although it is an example, does include normative keyword MAY.
Cyril: I think I understand the purpose of the change, but want to check.
… The last change, source -> implementation, I understand, that it says there's an implementation behind.
… Why was the word "single" removed?
Nigel: I don't think I can express the answer to that, it would be for Chris Wilson to say.
Cyril: Thanks, it looks good to me too. Will approve.
Nigel: Does anyone have any concerns about this pull request?
Pierre: This is your pull request though, Nigel?
Nigel: I did the admin of opening the pull request and typing the words, but the change was driven
… by Chris primarily, on the call.
Nigel: I'm sensing nobody has any concerns about this change.
… The next one is from Apple.
Nigel: Apple had a different take on it.
… They really want to push Charters to have a stronger requirement for CR exit than the minimum
… required by the Process, and I think in time they would like to raise that minimum in the Process too.
… So they're definitely unhappy about the idea that Content alone, even though it may have come from an
… implementation, as they see that, might be one of the factors.
… However, given that they have previously approved the old wording, they said if we have the
… old wording as well as the new wording, that would be hard for them to object to.
… They want to get W3C to a place where there are two implementations that do the intent of the spec,
… e.g. for a caption format, paints pixels on a screen.
Gary: They also understood our wording differently - what does "Content" mean,
… what does "validating implementation" mean. The old wording is tried and tested.
Cyril: Sorry if I'm rehashing, but the suggestion from Apple doesn't seem different from what we have.
… The Google change is clear about two independent implementation, not necessarily of the same type.
… The second thing: I'm not sure we're in disagreement.
… If I understand correctly it's W3T to verify implementations - they don't have to be open source,
… or on the market. That to me is equivalent to providing the content, because a tool has been used
… to provide it. I don't see how the proposal is any different to what we have.
Nigel: You reminded me of another key point of discussion:
… We haven't distinguished between content made for test by the WG and
… real world content made by a bunch of other implementations made by non-WG members.
Cyril: Chicken-and-egg - implementers want spec maturity before making content.
… [asks a q about content]
Nigel: I think a large cohort of real world IMSC content provided by multiple providers,
… all passing the IMSC HRM, would not be adequate from their perspective.
… They want the spec to be shown to be implementable with the same results based on the spec text
… by more than one person independently.
Pierre: I think it's bad to force Process changes via Charters.
… My interpretation is that they don't value the existence of documents as a means of proving interop.
… They want multiple processors independently made. That's the bottom line, right?
Nigel: Yes
Pierre: I don't agree with that personally.
Gary: Their issue with content is that to them it doesn't represent someone reading the spec and
… understanding it, necessarily. They could have thrown a bunch of angle brackets in a file
… and it just happened to work.
Pierre: I could argue the same about processors!
Gary: That was their issue with the validating implementation bullet.
… We had a higher model of what a validating implementation needs to be.
Pierre: My guess is if we accept Apple's input we will have that fight again later.
Gary: You're right but we likely have a better story for HRM particularly.
Andreas: Clarification: if you have an implementation built to create samples, that's an implementation?
… Also, about the normative keywords, there is no MUST in their PR right? It's a SHOULD.
Gary: Yes. They would prefer a MUST but are bringing back the status quo.
Andreas: Possibly that hints at what Pierre said that they may come back to this later.
Nigel: There are some options here:
… 1. Accept their PR
… 2. Reject their PR and tighten up the bullets to meet their needs
… 3. Wait for them to come back with potential alternative changes that would satisfy them.
Gary: They did say they would also circle back to the process discussions.
Nigel: Of those, does anyone think we accept their PR, so we can just move on?
Cyril: I need time to digest that.
… We've extended to when?
Gary: End of June
… If we don't come to a conclusion by then we'll have to extend again.
Pierre: Maybe to leave us, what do you and Gary recommend, having been in the discussions?
Nigel: Right now, personally, my recommendation is to think about it. The PR was opened 21 hours ago. Plus there may be other alternative options.
Gary: Right, there's no rush yet. Better to sit on it for a bit.
Pierre: Thank you
Meeting close
Nigel: Thanks everyone. Regrets from me for the next call.
… [adjourns meeting]