Timed Text Working Group Teleconference

<scribe> scribe: nigel

This Meeting

nigel: Any particular points to discuss, or other business not on the agenda?

group: No AOB.

nigel: Thanks all for attending at the new time - we will stick with this unless we know in
... advance that we need extra time, in which case we will begin earlier. Is that going to work for everyone?

group: no objections!

Andreas: That works for me.

IMSC

nigel: On IMSC 1.1, my summary is that we have established that we are able to publish a
... FPWD of IMSC 1.1 using the same short code as the existing IMSC spec, ttml-imsc1.
... We have it as a proposal to publish this.

Pierre: We have two pull requests that I'd like to accept today, and then I can generate a
... FPWD that can be used for publication.

Thierry: Then I will need to talk to the webmaster for publication on Tuesday at the earliest.

<tmichel> I confirm

Nigel: One small matter of admin for the record: http://www.w3.org/TR/ttml-imsc1/ has
... been updated in place today, to include a new link to the latest editor's draft.
... That was because the old link pointed to something in the github repo that we moved in
... preparation for IMSC 2.
... Philippe tells me that in April there will be a new plan to allow for a neutral ED link, which
... can be directed behind the scenes to the right place.

Thierry: Only the Rec has been updated - the Web master is not permitted to update any other
... version. We still have broken links in the old CR and PR. There is a GitHub hack that we
... can do.

Pierre: I don't think we really need to do. The Rec is the in-force specification that we need
... to worry about.

Thierry: Okay then let's not put energy into that.

Pierre: Also worth noting, GitHub's gh-pages is now able to publish from any named branch
... so I set it to publish the ED from the master branch. That closed issue #187 https://github.com/w3c/imsc/issues/187

Nigel: Great, thanks for that!
... That means we are serving everything on the master branch but I don't think that is a problem.
... Let's look at the pull requests then.

-> https://github.com/w3c/imsc/pull/205 Stub to support the current ED link found in the IMSC1 REC

scribe: Now we have updated the Rec we don't need this.

Pierre: I just closed that pull request and the associated issue #204.

-> https://github.com/w3c/imsc/pull/200 Add support for fillLineGap and optional features

Pierre: I think everyone was okay with the text. I made a modification to change the example
... to the nice one that Andreas provided.

Nigel: I just want to check if we really consider it necessary to have an example with
... multiple font sizes on a single line?

Pierre: It is non obvious how it should work, so since the feature works in that case we really
... ought to show what the intended effect is, or remove that possibility.

Andreas: I agree. With the text as it is now we should include this level of complexity in the example.
... I also would be happy to further constrain it.

Nigel: I'm happy with the feature, I'm just not sure if the example and/or the test case needs
... this level of complexity.

Pierre: We should because it is allowed.

Andreas: I agree - there are other examples that are more simple than they could be, and that
... can cause interoperability problems later.

Nigel: I think it would be easier to implement as is than to specify the constraint that would make it easier to implement!
... I would prefer to leave the text as it is now.

Andreas: This is an interesting discussion, which we should continue after publication -
... and can do into WD.

Pierre: Something Nigel said last time is we can go back to the requesters for the feature
... based on the FPWD and check that they are happy with the level of implementation complexity.

Andreas: Makes sense to me.

Pierre: That's an argument for exposing that complexity in the FPWD.

Nigel: That's fair.

Pierre: Long term, we really need to get the CSS folks involved in this.
... Again, with a FPWD it will be easier to go back to the CSS WG and ask what to do about this.
... Either it is essential for the web and should be in CSS eventually, or its not.
... I propose that we merge this PR.

Nigel: Any other comments on this before we hit the big green button?

group: No.

Nigel: Okay Pierre, go for it!

Pierre: Do you want to add a note?

Nigel: Ok, I've done it.
... That's also closed https://github.com/w3c/imsc/issues/195
... Next pull request:

-> https://github.com/w3c/imsc/pull/202 Clarify relationship with current REC

Pierre: review the last commit at https://github.com/w3c/imsc/pull/202/commits/d348dc063baf9395df7bc1947a5dd886926c0c87

Nigel: Ok, any comments on this, requests for changes etc?

group: No comments.

Nigel: Okay, I'll comment on it and merge... Done!
... That's also closed https://github.com/w3c/imsc/issues/201

Pierre: There are no issues open.

Nigel: All 12 open issues are labelled IMSC2 milestone.
... Just checking the latest ED... http://w3c.github.io/imsc/imsc1/spec/ttml-ww-profiles.html#itts-fillLineGap
... The image is not being served.
... The case of the PNG extension differs in the repo vs the link.

Pierre: I'll rename the file.

Nigel: OK that's fixed now.
... We have a proposal to publish a FPWD based on this ED.

Pierre: I'll create a branch with the SOTD updated to suit and point the group and Thierry to it.

<tmichel> OK.

RESOLUTION: Publish a FPWD of IMSC 1.0.1 based on the current ED.

Nigel: Great, thanks everyone.
... Has there been anything on the CLDR Unicode liaison?

Pierre: No.

TTML

Nigel: I want to raise Security and Privacy.

-> https://github.com/w3c/ttml2/issues/224

Nigel: You might also find interesting a parallel piece of work, for reference:

-> https://github.com/w3c/webvtt/issues/323

Nigel: I would like to encourage everyone to think about any security or privacy considerations
... associated with TTML2 that we should declare but that we have not already captured
... in the ttml2 issue 224 notes.

Pierre: I think that padding on span in TTML2 as currently specified does not do what we
... want it to do, for ebutts:linePadding. Padding on span does not accomplish that.
... Note that box-decoration-break: clone; does not work with nested spans and
... multiple colors.

Nigel: Oh yes, because if you have two adjacent spans with different background colors then
... you want the padding if they happen to have a line break between them but not otherwise.

Pierre: I see no value to adding padding on spans; if linePadding needs to be supported then
... we need to file an issue.

Andreas: The question is if we should go beyond what we have in EBU-TT-D. It is a similar
... problem with line gap - we define the presentation semantics without a clear view of
... how to get there. If there is no structural correspondence in CSS to solve this then we
... may just want to define the presentation semantics, or possibly it is too complex to fully
... implement it.

Pierre: I would put this in the "work with CSSWG" bucket. I'm not sure where we will end up
... but being clear in the next working draft that this is a feature that the group intends to
... make part of TTML2 and is essential and working with CSS WG, otherwise it will be poorly
... supported. My suspicion is most people won't do it correctly.
... I'm not questioning the usefulness of line padding, but just want to point out that there's
... a missing feature in TTML2, which will make it hard for IMSC2 to be a subset of TTML2.

Nigel: Will you raise an issue?

Pierre: Ok I will.
... Next subtle point is multiRowAlign, which if implemented in CSS in the same way as
... hinted at in TTML2 does not actually work, because browsers handle lines with explicit
... br breaks differently from those split by word wrapping. This is subtle! I think we again
... need to work with CSS WG to show how important this is and how it can be fixed.
... In the meantime I will file an issue.

Dae: If these are important issues we should try to have a joint meeting with them.

Pierre: We may be able to request time in their April meeting in Tokyo.

Nigel: The first step is for me as Chair to reach out and discover if we can get some agenda
... time with them to discuss these issues.
... So I'll do that. We're out of time for today, so thanks all! [adjourns meeting]