Timed Text Working Group Teleconference

<scribe> scribe: nigel

This meeting

Nigel: Today we have IMSC 1.0.1 and TTML issues. Any other business?

Pierre: I would like to review our progress against goals for TTML2, issues closed etc.
... I would like to check we are still on target and see if there are any blockers.

group: No other business

IMSC

nigel: Since we last met we have published the FPWD of IMSC 1.0.1 with a new short name.
... The old link now redirects to the FPWD which has a link to the Rec.

Thierry: The publication was a bit chaotic because the original agreement to use the same
... short name was in fact not possible and the web master refused it.

nigel: My concern is that anyone using the /TR/ttml-imsc1/ link no longer gets to the Rec
... but now goes to the FPWD. To link to the Rec we now have a special dated link.
... I would propose that we remove the redirect from ttml-imsc1 to 1.0.1 and reinstate it when
... we have published 1.0.1 as a Rec.

Pierre: That works for me.

Thierry: I can ask for that. I propose to add something:
... 1. Undo the link from ttml-imsc1 to ttml-imsc1.0.1
... 2. Remove the Latest Rec link from 1.0.1

Nigel: No we should keep that

Pierre: That's always useful.

Thierry: We could also change the name separately from the short name, but that would be confusing.

Glenn: No they need to be connected together. I do object to using 1.0.1.

Thierry: We had that conversation Glenn, I think you were not present.

Glenn: I did get my comment in within the Decision Policy review period.

nigel: So just to confirm Thierry will ask to remove the redirect from ttml-imsc1 to ttml-imsc1.0.1 for now, and then when we have published 1.0.1 as a Rec reinstate it.

Pierre: Is this the Wide Review version?

nigel: It can be.

Thierry: If you want to add a date for Wide Review we can publish an updated WD as soon as we like.

Pierre: So we need to add a deadline to the comments and request specific review of the
... added features.

Thierry: I would name those features.

Pierre: Yes of course. Should I also add those to the SOTD?

Thierry: I think so.

nigel: We should deal with the typo in pull request

=> https://github.com/w3c/imsc/pull/207

nigel: while we are doing a change.

Pierre: Looks good to me.

Nigel: I can send liaisons out requesting wide review to the usual suspects, probably not
... until early next week. I will draft the text and circulate it.

Thierry: I will take care of the internal W3C horizontal review requests.

Nigel: Great, thanks that's a good division of labour!

Pierre: In that pull request I factor out the file as external to make it easier.

Nigel: Does that mean you can validate too?

Pierre: Yes exactly!

Nigel: We don't have an updated schema for 1.0.1 do we?

Pierre: That's a really good point! Thanks for reminding me, please can you file an issue?

Nigel: done:

-> https://github.com/w3c/imsc/issues/208

Nigel: For a SOTD review deadline I'd propose Friday 31st March.

Pierre: Works for me.

Nigel: To summarise the actions:
... NM to draft external liaison message;
... PAL to update WD
... TM to initiate Horizontal Review
... TM to request temporary removal of redirect
... PAL to propose a fix for the schema issue

TTML

Nigel: We have total 67 issues open on TTML2, of which one is marked as for TTML.next
... 6 issues were closed this week.
... We said that we would have a WD out by end of March, so how are we doing Glenn?

Glenn: Given my schedule I'm confident we will make that and there are going to be slews
... of updates over the next couple of weeks. It will depend also on how much conversation
... we need to close these before we're ready to call them ready for a WR draft.
... I'm hoping we don't have too many new issues filed.

Pierre: We set ourselves a self-imposed deadline of Feb 15 for new feature requests - are
... we sticking to that?

Glenn: Yes, unless to resolve an issue we need new features.

Pierre: I think any feature requests after Feb 15 we would defer.

Glenn: I agree. I don't have anything on my list that would fit into the category of a new feature.

Nigel: That works for me.

Pierre: Then in order to get consensus at the end of March we need all the PRs done 2 weeks
... before the end of March to give us 2 weeks to accept all the PRs.

Glenn: That's a good goal that I'll endeavour to achieve.

Nigel: It's going to be a big review task - a good one, but a big one - so if we are later than this then we should delay WR publication.

Pierre: Just to be clear, it won't be acceptable to me if we want to publish a WR draft only a short period
... after the PR has been merged.

Glenn: In that case I will try only to push editorial changes out beyond the deadline rather than substantive things.
... If there are any controversial issues please could you let me know so I can give more time for review?

Nigel: Just one from me - the Audio Description features are a departure, so we should
... aim to do those sooner if possible. I also have a presentation at CSUN at the beginning of
... March during which I would love to be able to reference a document.

Glenn: I aim to do that next week.
... Regarding HDR, do you anticipate any changes?

Nigel: Setting a cd/m2 level is controversial - we should talk about that.

Pierre: Happy to discuss that.

Nigel: Okay let me get the right people in the room - I'm not trying to exclude the proposal
... where it's needed, but to ensure that for all HDR schemes we have a working solution.
... Okay that's planning stuff done, any issues or pull requests that anyone wants to discuss?

Pierre: I'd like to go back to textEmphasis and the Mongolian script divergence.

-> https://github.com/w3c/ttml2/issues/240

Glenn: There's no technical issue here - it's just a matter of tokens not semantics.

Pierre: So if I look at before and after and compare them to over and under there is a 1:1 mapping?

Glenn: It depends on writing mode.
... I had a question about text orientation and whether it is sensitive to that, and I convinced
... myself that it is not, and is only dependent on writing mode.
... Also I do not have enough context to answer the question - it's important to know if we are
... talking about vertical text embedded in horizontal text in an inline block, or just vertical
... Mongolian. I also am not aware of any use of text emphasis in Mongolian.
... Most Mongolian writing is in Cyrillic; there are some usages in Inner Mongolia where
... it is written with Chinese, but generally Mongolian vertical script is not very common.

Pierre: I'm concerned that Elika thinks that the CSS system is different, and want to check that they do in fact align.

Glenn: We might defer this to TTML.next in case some substantive point is uncovered.

Pierre: Moving on, the auto value for text emphasis is not supported in CSS. Elika stated
... that's never going to happen in CSS, and I pointed out that it could be used for captions
... so she asked for some examples and use cases. Either we should point that out or consider
... excluding it from TTML2 or removing it as the default.

Glenn: "auto" came up when we introduced "outside", and that turned out to be the default
... scenario in the popular Japanese subtitling system. It only really applies if there are two
... lines so I had to come up with semantics for 1 line or >2 lines. That's where that came from.

Pierre: The problem is the way "outside" is specified it applies a different semantic depending
... on the number of lines and whether it is the first or last line. In CSS it's really awkward to
... do that because you have to figure out where the line breaks are.

Glenn: I agree you have to discover where the line breaks are and treat each line separately.
... The fact that TTML2 and CSS differ is not a problem per se.

Pierre: It is for Movielabs - it will generate an objection unless there's a fallback or if there's
... an issue filed on CSS. The door is open to provide the use cases and file an issue or specify the fallback.

Glenn: I don't see a problem with doing both.
... We haven't got consensus that CSS mapping is mandatory.

Pierre: Folks interested in IMSC 2 are very interested in the mapping to HTML and CSS and that's really important to MovieLabs.

Glenn: Would you be open to defining the fallback in IMSC 2?

Pierre: Absolutely, that sounds like the right place to do it, but we have a catch-22 here
... because we're working on both in parallel.

Glenn: We have some fallbacks in TTML and IMSC already. We don't have any CSS fallbacks though.
... Opening that door could be problematic because it might require treating everything in TTML2
... with a formal mapping to CSS.

Pierre: I don't disagree with that, I'm just concerned that after TTML2 is in CR we find a
... fundamental issue in defining a fallback for IMSC 2 so I'd like to keep track of the fallback
... while we are working on them both.

Nigel: By the way, if we need to be normatively dependent on CSS we will be stuck for a while
... until all the relevant CSS specs have reached Rec.

Pierre: I just wanted to make clear that this is really important for Movielabs.

Nigel: Ok we're out of time, we have some actions and discussions to have offline and
... will be back here on the phone next week. Thanks everyone. [adjourns meeting]