Meeting minutes
This meeting
Nigel: Today we have IMSC 1.2 ARIB Liaison response, TTML2 bits and bobs, and TPAC 2020 planning.
… Any other business, or points to make sure we cover?
group: [no other business]
IMSC 1.2
Nigel: The AC Review is continuing, and I encourage everyone to make sure their AC rep
… votes!
ARIB liaison response w3c/ttwg#116
Nigel: I've begun drafting!
… [shares screen with early preview]
Nigel: Mike, I was wondering if your response on https://github.com/w3c/imsc/issues/547#issuecomment-640801013
… is that to the question at https://github.com/w3c/imsc/issues/546#issuecomment-638954715
… Since they're two different but related issues.
Mike: I don't think so, I'll research that right now. I think I sent a private comment not on
… the GitHub issue. [checks]
Nigel: Ah, then there may be some extra information I may need to fold in.
Cyril: I wonder if there should be a sentence after the issue list, because I'm not sure,
… reading quickly, in which direction this is going. What is the summary of their comment?
… Should we say something like "some comments are accepted as is, others need more
… information"?
Nigel: Good point.
Cyril: [suggests adjustments to how we ask for usage information]
Mike: More information about the response I got about images just sent to the TT public reflector.
Nigel: Thanks for that.
… The last point, about #550, I was slightly less clear about.
Cyril: Re GSUB, PUA and IVS, we said PUA may not be the best method because of
… degradation not being graceful.
Nigel: We haven't discussed the "end game" of potential convergence between standards,
… and if ARIB might one day want to adopt IMSC if it has the right feature set, but I think
… our idea is to assess the differences and if we can come to a better understanding
… before moving to that.
Mike: Shouldn't we find a polite way to suggest that they participate if they are interested
… in convergence to a single standard?
Nigel: Interesting point, yes, I haven't considered that too much.
… Any other views?
Mike: Or we could ask here if anyone wants to work on it without an ARIB person present.
Pierre: I think there are still unanswered questions in my mind on which we can make
… progress without direct participation in TTWG. For example it would be good to hear
… their opinions on negative letter spacing, PUA, etc.
… This is information exchange that might make IMSC better.
… But yes, before asking about switching to IMSC, that is a bigger discussion that
… really requires face to face or private discussions.
Mike: To the extent that we can make TTML2 and IMSC better and more palatable to the
… Japanese TV market, that's goodness, but trying to orchestrate something larger and
… addressing all these issues, I don't think we can make progress without active
… participation of someone with a detailed knowledge, first hand.
Pierre: I would keep this liaison to the specific questions or seeking specific information.
… At this point, PUA, IVS and negative letter spacing really come to mind. Examples are
… really critical.
Nigel: I would keep the whole list rather than pruning it, I'm interested in audio certainly.
… And animation might be a relatively easy thing to specify, based on TTML2.
Pierre: Maybe some of these questions arise when we get closer to alignment.
Mike: Maybe we should make it clear early in the message that we invite ARIB members
… to participate in W3C, and to contact the Chair. If we got a good engineer we should
… figure out how to let them participate, regardless of existing membership.
… One other point: there's an English version version - did you get that?
Nigel: We did see a link to that in the past.
Mike: I could put a link across the reflector.
Nigel: No harm repeating it!
Mike: Will do.
Nigel: Thank you all very much for the input, that's helpful. I'll find a way to share this
… in a private space for further review.
TTML2 2nd Edition
Nigel: You should have seen that we now have a date to discuss our PING issue, which is
… Monday - there's an invitation and details on the member-tt reflector.
… On the TTML2 Implementation Report, I saw some updates from Glenn, so some features
… have an implementation. I am not aware of any other implementations yet, especially
… for the validation functionality.
Virtual TPAC 2020 Planning
Nigel: The Chairs have been asked to plan for virtual joint meetings under the "virtual TPAC"
… banner. So far I think we would like to talk to:
… MEIG
… WebMedia WG
… CSS WG
… Any others? Or should we try to merge MEIG and WebMedia WG for example?
Cyril: Is WebMedia WG the Media WG? I'm not sure there's a "Web" in front.
Nigel: Apologies, I got the name wrong.
Gary: Yes, it's just Media WG
Nigel: Thanks for the correction!
Cyril: I agree a 3 way meeting is probably better.
Gary: Yes unless there's something specific for one but not the other but I highly doubt it.
Nigel: I'm not aware of anything like that and I think a lot of the people are the same.
Gary: Definitely a large overlap between those three groups.
Cyril: What is the status of the Text Track v2 work, Pierre?
Gary: It's in Safari
Pierre: I don't know, but apparently it's in Safari!
Gary: I guess they shipped it as an experimental feature.
Pierre: I'm "pleasantly surprised"! [laughter]
… I thought they would go down the path of HTML not a bespoke API.
… Do you have a link to the announcement?
<gkatsev> safari 13.1 announcement
Gary: They mention this is available in the Safari 13.1 announcement
… That links to an explainer that apparently no longer exists.
… They renamed the branch name. Here's the explainer:
<gkatsev> texttrack explainer link
Gary: You provide a specially formatted HTML fragment. That's available as an experimental
… feature in Safari.
Pierre: It's not the original idea of having an API specific to timed text, it's a more generic
… HTML based approach. 7 years ago this was my original feedback, why not allow
… general HTML to be rendered. They said "security issues..." but maybe they've
… decided that it is okay.
… Thanks for pointing that out!
Nigel: I'm super excited to read about this.
Pierre: My name is everywhere here, too!
Gary: I noticed because we had a bug in video.js for captions and spotted the announcements.
Pierre: On a side note, I guess this is webkit so if I was subscribed to webkit GitHub I would
… get those announcements.
Gary: I subscribed to the webkit blog.
Pierre: Okay. thanks a lot for pointing that out.
… I guess from an IMSC perspective what it really does is...
… We still have to handle client side conversion of TTML to HTML but all the timing
… is handled by this.
Gary: Yes, it sounds like there are a couple of special attributes to specify the cue part
… and the cue background, but otherwise it's HTML.
Pierre: I get it, so they followed the TTML model but as HTML, a subset that you can use
… that happens to be [scribe missed]
… The explainer is 6 months old!
Gary: Yes, I believe this shipped in mid-March.
Cyril: Do you know if other browsers are adopting it?
… It leads to nowhere if it is only in Safari.
Gary: Yes, I think Google is the one that needs convincing.
Pierre: My recollection (7 years old, maybe exaggerated) was this idea of passing HTML
… to a cue was a terrible idea from a browser perspective, but if the model were constrained
… that would make it easier.
Nigel: I remember the same thing.
Pierre: I will follow up with Eric and Tess offline.
Gary: I think because they limit the allowed elements it makes it relatively safe.
Pierre: Right, and even though the explainer doesn't say it, it could be that the
… relationship between the elements may be further constrained. The early implementation
… used JSON and enforced the TTML data model, like a p has to be within a div.
… I wonder if they still enforce that, because if so, it further reduces the risk.
Gary: I'm not sure about that.
Pierre: It's not in the explainer.
Nigel: Back to the topic, there's an action for the Chairs and staff to arrange those meetings.
… So far no other groups have been requested.
Nigel: As an aside, CSS WG has an upcoming "virtual f2f" which has an issue on the agenda
… for considering the name of their equivalent of the fillLineGap property, and it's possible
… they will revisit the background painting area property for linePadding too, and bring them
… together.
[css-inline] inline-sizing property name is too similar to inline-size w3c/csswg-drafts#5189
Meeting close
Nigel: Thanks for the stimulating meeting everyone. Apologies we're 3 minutes over. [adjourns meeting]