Meeting minutes
This meeting
Nigel: Today we have IMSC 1.2 transition request, TTML2 IR, and TPAC planning
… None of them are big topics.
Nigel: Any other business, or points to make sure we cover?
… I can summarise if anyone wants about the font fingerprinting TTML2 issue, in AOB.
group: [no other business]
IMSC 1.2 transition request
Nigel: Atsushi, thank you for raising the transition request
… No objections in the AC review
… The change feedback has already been addressed.
Nigel: One question on IMSC 1.2 Rec I think we have not answered yet is:
… Obsolete IMSC 1.1 or leave as a Rec also?
Pierre: I think IMSC 1.1 remains in active use and should not be obsoleted with 1.2.
Nigel: I think that's right.
Atsushi: Obsolete or superseded means that in e.g. IMSC 1.1 will be marked as inactive
… by IMSC 1.2, but I think that is not the case for this time.
Nigel: Sounds like we're all in agreement. Any counter-views?
group: [silence]
Mike: Not a counter-view, but a question. 1.0.1 remains in use and people are referencing
… it. What are you proposing with IMSC 1.1?
Nigel: Not proposing anything, just checking in since we didn't discuss it.
… This is the easiest time to obsolete the previous version, but I don't believe we want to.
… Thought I should check in and confirm my understanding/expectation.
Mike: I understand, thank you.
Resolution: On publishing IMSC 1.2 do not obsolete or supersede any previous version.
Nigel: Just for clarity, we don't need a Resolution to request transition to Rec because it
… is not for us to do.
Atsushi: That's right. If we don't decide to do anything we will have 3 active IMSC specifications.
Nigel: Good, I notice that Pierre has already prepared the Pull Request for Rec, thank you,
… and Atsushi and I have already approved it.
… It may need a tweak to the date, which can be done later easily.
… Any other points on transition to Rec?
Atsushi: No, not from me.
ARIB liaison response
Nigel: No progress from me after discussion previously.
… I did think about how to work on this in a private space, and the only option I am
… aware of is the member-tt reflector, which doesn't feel like a great way to iterate over
… a document. Any other ideas?
Pierre: What about creating a google doc and just sharing the link on member-tt?
Nigel: Yes, that'd work!
<atsushi> +1
Nigel: OK, fine, will do.
TTML2 2nd Edition
Nigel: Glenn has made some updates recently, indicating one passing implementation
… for many of the tests, but not yet all of them.
Nigel: I don't know of anyone planning a second implementation for these tests at the
… current time. We may find ourselves in CR for a long time, if nobody steps forward.
Cyril: I checked the exit criteria again. I wasn't clear if we need a validation or presentation
… or both. I read that at least two implementations must support each changed feature.
… So it doesn't matter if it is presentation or validation?
Nigel: Right
Pierre: The IR is up to date?
Nigel: I believe so
Pierre: It doesn't look so great. None of those are at risk, right?
Cyril: We have nothing at risk at the moment.
Nigel: I don't think we added any new features in 2nd Ed, as such.
… So we maybe need to consider exactly what we think is needed for CR exit a bit more
… more carefully.
Pierre: For instance, all the presentation tests are clarifications, not new features?
Nigel: Arguably audio-in-body is a new feature.
Pierre: For that we do have an implementation.
… I'm looking at all the other ones.
… We should try to make sure the page reflects reality. If these are not new features and
… are just clarifications... In IMSC, we've added tests with clarifications without necessarily
… having a new round of the IR. Maybe we shouldn't be so hard on ourselves!
Nigel: I have sympathy with this, and will try to assess each test and its related change
… a little more carefully. Counting against that in my view, the spec should reflect implementation
… reality generally where possible.
Nigel: My understanding is that Glenn plans to get around to adding code to TTT to
… pass the remaining tests.
Pierre: Also, and we've discussed this before, where one is an open source implementation,
… where one is a reference, it is a lot less important that there are 2 independent
… implementations.
Nigel: I remain unsure about that!
TTML2 Font fingerprinting privacy issue (summary)
Nigel: The minutes are already sent out from Monday's meeting.
… In essence, we should try to get as close to the request as we can, and make the
… transition request when we're ready.
… My understanding is that this will be considered a reasonable approach.
group: [requests non-scribed discussion]
Pierre: Is there a deadline for submitting the pull request?
Nigel: It's our timescale.
Pierre: We ought to let Sam review it in a bounded timeframe.
Nigel: My view is that we treat this like any other pull request, merge, proceed, and tell
… Sam about it when it's done.
Pierre: Thanks, that answers my question.
Virtual TPAC 2020 planning
Nigel: Thank you Atsushi for making requests for joint meetings with MEIG + Media WG,
… and CSS WG.
Atsushi: I'm waiting for their reply. If we have any preference on day, e.g. Monday/Tuesday,
… it may be possible to propose something to them just for that.
Nigel: Can I ask you to send a reminder to TTWG please, summarising the virtual TPAC
… timelines and dates etc?
Atsushi: Yes, I will send that.
Nigel: I think we have specific topics for MEIG and Media WG, but not for CSS.
… Please could I ask you all to think about which CSS issues need joint discussion?
… By the way. CSS WG is having a virtual f2f this week, and immediately before this call
… I joined, because there was one scheduled TTML issue on the agenda, which is the
… name of the fillLineGap-equivalent property. That discussion did not conclude yet.
[css-inline] inline-sizing property name is too similar to inline-size w3c/csswg-drafts#5189
Nigel: Possibly the shear issues for example need more discussion between us and CSS WG?
Cyril: It depends on how the shear issue evolves until TPAC.
Nigel: Fair enough!
Cyril: Is CSS WG working on overhang also?
Nigel: I don't know.
Cyril: I think that was one issue they wanted to work on at some point, maybe deferred until the next version.
Pierre: What's the best way to work with CSS? If we need changes in CSS we need a concerted
… effort. Is this the right time to do it?
Cyril: I can just repeat what I said on the email thread.
… Several ways to work with CSS WG. Japanese subtitles are my main concern.
… Either go through Japanese Layout TF, who are focused on print, and harder to convince
… on subtitles-related issues.
… I'd rather persuade CSS WG directly that is is important, keeping JL TF in the loop.
… Best way is to get the feature implemented in browsers I think, then they will talk in CSS WG
… about the best way to do it.
Pierre: It's important that subtitling community people are included in prioritisation discussions about styling that affects subtitles.
Atsushi: There is a lot of work to do, and many complex interactions across other CSS properties, e.g. tate chu yoko and oblique together.
… Even for now, applying underline to Japanese in horizontal writing mode is quite complex.
… For a wider variety of decoration, a question for JL TF and CSS WG is all the permutations that
… should be supported and their behaviours.
… For now Chrome's implementation allows tate chu yoko to be expanded to 1.1em, but
… that could conflict with several decoration styles, so if we can consider all of these legal (illegal?)
… combinations it might be possible to go forward. But it is quite difficult to cover them all.
… For manual typesetting editors will coordinate everything but for CSS we need the capability
… for all the decorations. All of these special cases need to be considered for the specification.
… That is the question actually, and which ones may not be needed for timed text or subtitles.
Meeting Close
Nigel: Apologies for going over time, very stimulating discussions. Thank you! [adjourns meeting]