See also: IRC log
<scribe> scribe: nigel
Nigel: First, welcome Cyril, who
has joined this group (again), now representing Netflix.
... For today, we have TPAC agenda, TTML2 CSS mapping
semantics, TTML2 WR comment disposition process.
... Any more?
Thierry: For WebVTT, just to note
that the Editor, Simon, has stepped down, so we are
... lacking an editor.
Nigel: I had the impression that Sylvia had offered to be a second Editor - is that not correct?
Thierry: She used to be an
Editor, I'm not sure if she will come back to that former
task.
... Also just a small reminder for people to register for TPAC
if not done, and of course to
... invite new people like Cyril to join the face to face at
TPAC.
Cyril: I plan to attend!
Nigel: The first question is:
does anyone plan to join remotely?
... If so please let me know by October 20th at the absolute
latest, otherwise we will not
... have a speaker phone.
... If you have not yet properly registered please do so soon
otherwise it will cost more to
... do it later.
action-504?
<trackbot> action-504 -- Nigel Megitt to Send csswg an invitation reminder by 2017-09-07 -- due 2017-09-07 -- OPEN
<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/504
Nigel: I haven't yet done this
[slaps wrist] but will do so before the end of the week.
... As well as the joint meetings there is also the proposal
for a demo slot on the
... Wednesday, which Pierre and I are in the process of working
on.
... [iterates through draft agenda topics]
Pierre: Can we schedule IMSC 2 requirements review for Thursday afternoon please?
Nigel: OK noted
Pierre: The goal should be to publish the requirements following that meeting.
Nigel: +1
Pierre: (unless we have done it before)
Nigel: Anything else to add to the topics list?
Group: [nothing else to add at top level]
Nigel: Anything to add as detailed points?
Group: [nothing else to add as further detail]
Nigel: On TTML <--> WebVTT mapping or HTMLCue, Andreas is there anything to discuss?
Andreas: I'm not sure the group
is really working on HTMLCue, so there probably won't
... be anything.
... For the TTML <--> WebVTT mapping we may spend a bit
of time. I think we said it
... makes no sense to do the mapping if the WebVTT spec is not
stable, so we can assess
... the status and discuss how to proceed. That's possibly the
best we can do. Maybe we
... can take 15-20 minutes on that, it should be enough.
Nigel: Okay, thank you.
... Okay that's a good quick review of the topics. Now in terms
of scheduling discussions,
... we have CSS joint meeting on the Friday - I guess it makes
sense to do that in the morning.
Pierre: There will be a public
holiday on the Saturday that is often observed on the
Friday.
... I doubt it will impact our meeting much.
Nigel: We'll fit the other things
in around any requests or constraints - if there are any
... then please let me know.
... Any other points to raise on TPAC?
action-506?
<trackbot> action-506 -- Thierry Michel to Draft a wiki page explaining our review and disposition steps and labels -- due 2017-09-14 -- OPEN
<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/506
Thierry: I did this, and sent an email to the group.
TTML2 Wide Review comment wiki page
Thierry: We have comments either
coming on the public mailing list or directly on github.
... Mainly HR like i18n comes on GitHub and the public sends to
the public list.
... In order to process those comments we discussed adding
labels on the GitHub repo,
... so I went through the different steps. The first category
is processing comments within
... the group, getting to a full or partial resolution, and
mentioning what type of comments
... we are addressing, substantive, editorial, typo, general or
undefined.
... Then the second step is to get approval from the commenter
or if there's no agreement
... go back to discussion on the first steps.
... I have one issue I wanted to discuss. When we are resolving
a response to the commenter,
... either fully or partially, should we first edit the draft
and then send the response or wait
... for approval before modifying the spec. Glenn, that would
be an action on you I guess, so maybe you have a comment?
Glenn: I think we do not need to
consult with the commenter before taking action. It is up
... to us if we want to, or to ask for clarification if we want
to.
Thierry: This is new, to use GitHub rather than the disposition of comments tool.
Glenn: I think there's only one
label right now, called Wide Review comment, unless we
... also consider the HR comments, which I would lump in.
Nigel: Actually the Process puts HR comments in as part of WR.
Glenn: Okay
Nigel: We have these
labels:
... wide review comment
... horizontal review comment
... wg accepted
... needs discussion
... editorial
... awaiting disposition response
... bug
... enhancement
... It looks like some are for comment type categorisation -
editorial/bug/enhancement
... And others map more clearly to Thierry's steps.
... Thierry it would be helpful if you could look at the labels
and see if there's a mapping,
... and if not, propose any new ones needed.
Thierry: Ok
Current TTML2 repo issue labels
Thierry: Another clarification.
For the first category - if we discuss a comment we
should
... conclude the discussion with a label that the WG assigns.
For the second category,
... which is more process, I will take care of that, drafting
and sending a response and adding
... the appropriate labels.
Nigel: Just to be clear, the
first category is "WG processing comment" and the second
is
... "Sending disposition back to the commenter for
approval"?
Thierry: Exactly, yes.
Nigel: And then the idea is we
can use the GitHub issue labels to provide the right
information
... for Transition?
Thierry: Yes, so we can justify
to the Director how we dealt with each comment.
... Glenn, so when we have a resolution within the group, at
that stage do you want to do
... the spec update?
Glenn: Yes, although in some
cases we don't have a formal resolution that requires an
editing
... action. Sometimes I do an edit that the group then accepts.
I could go both ways.
... Technically for editorial changes, I don't even have to go
through a review period.
... The degree of formality may vary depending on the
issue.
Nigel: I'd add another step in 2, which is "No feedback received from commenter within the stated period".
Thierry: I will add that.
Nigel: Formally we don't need to
have a dependency between spec updates and sending
... dispositions to the commenter - we just have to have
everything complete by the time
... we request transition.
action-506: [Meeting 2017-09-14] Need to add GitHub labels and extra 2.7 step as discussed
<trackbot> Notes added to action-506 Draft a wiki page explaining our review and disposition steps and labels.
action-503?
<trackbot> action-503 -- Thierry Michel to Send a reminder to groups for ttml2 review by 8th september -- due 2017-08-17 -- CLOSED
<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/503
github: https://github.com/w3c/ttml2/issues/406
Nigel: When we discussed this
before I took the action to create actionable specific
changes.
... I did that, creating a bunch of sub-issues, and that
generated some further offline discussion.
... The result is that I will prepare a PR putting the semantic
basis informatively into a new
... appendix, for each style attribute, and link from a new row
in the style attribute table
... to the relevant place in that appendix. Within the appendix
I will include the XSL semantic
... and the equivalent CSS semantic.
... That will remove a Note from every style attribute. Before
we go into the detail, does
... that sound okay to everyone?
Glenn: That sounds okay to me.
Right now we have an informative note in each of the
... style sections that point to a particular XSL-FO property,
and we have a derivation appendix
... too, so you should consider moving that information into
the new appendix, or leaving
... it intact and updating it. We don't want to end up with two
different incompatible sets
... of derivation information.
... If adding these new references to CSS references results in
any new documents being
... added to the reference section then they should go into the
informative references not
... the normative ones. That way I am less concerned about the
status of the CSS documents.
Nigel: One thing to note from the
TAG meeting I attended was that timbl pointed out that
... W3C has agreed that Recs can normatively reference CR
documents, so there should be
... no concern about doing that normatively, at least in
principle.
Glenn: Many of the CSS documents are not yet in CR.
Nigel: That's a good point.
... Moving into the detail, I've listed in each sub-issue the
XSL reference and the proposed
... CSS equivalence, so I'd really appreciate any comments on
those before I do the editing.
... I'd like to mention a couple straight away.
... The first one is that the textDecoration mapping seems like
it might be quite complicated.
... This is because of our inheritance rules for
textDecoration, and because XSL 1.1 makes
... differences compared to the version of CSS that it
references, and it's not clear to me yet
... if a later CSS version has adopted the same changes.
... I will raise an issue for textDecoration, as well as
visibility and zIndex, to complete the
... set and allow a place for discussion.
... By the way, I only raised issues for the style attributes
that currently reference XSL.
... All the others either don't reference anything or reference
CSS already.
Glenn: On that point, we should
be consistent with the use of this appendix so that all
... references are vectored through it, or as many as possible
that have the same basic
... structure.
Nigel: Agreed.
github: https://github.com/w3c/ttml2/issues/415
Nigel: I wondered why this uses
flex-direction: column and then justify-content, rather
... than the normal inline direction and align-content?
Pierre: I did do some playing
around with this. align-content doesn't behave in the
... expected way, and only applies for multi-line content, and
also that column direction
... means that the before and after edges align properly. I can
add notes to the issue.
Nigel: That would be great, thank
you.
... Is tts:displayAlign writing mode dependent?
Glenn: I think of it in
writing-mode relative terms, so if writing mode is tbrl then
the
... alignment axis would map to horizontal in the absolute
sense, so "before" would align
... to the right.
Nigel: I wonder if we have any tests for that?
Glenn: I know that ttt has both
syntactic and semantic tests for that but I'm not sure if
the
... TTML1 original test suite had them, it may not. It probably
has something for horizontal
... writing modes only, I'd have to check.
Pierre: I'm looking at the IMSC1
tests... there are a lot of writing mode tests - yes,
there
... are some tests for displayAlign there, including one that
looks like it was designed
... specifically to trigger that - writingMode10.
<pal> https://github.com/w3c/imsc-tests/blob/master/imsc1/ttml/writingMode/writingMode010.ttml
Pierre: It was actually in response to ttml2#269, so maybe there's something there.
Nigel: Since imscjs passes the test, I'm a lot more confident in saying this is a good mapping.
Glenn: I just checked and none of the TTML1 tests check this case.
Pierre: writingMode10 was added
on top. Also some of the IRT tests include a combination
... of writingMode and displayAlign, and they're also included
in imsc-tests.
<pal> https://github.com/w3c/imsc-tests/blob/master/imsc1/ttml/writingMode/writing-mode-tbrl-001.ttml
<pal> https://github.com/w3c/imsc-tests/blob/master/imsc1/png/writing-mode-tbrl-001/2.000000.png
Nigel: Great, that makes sense, thank you!
Pierre: The latest release version of dash.js now has support for IMSC 1 Text and Image.
Cyril: As a side file or as embedded MP4?
Pierre: It should do both.
Cyril: Great!
Nigel: Good news indeed.
Mike: I'm unclear that's true -
certainly the sidecar works fine, but last time I looked at
it
... it was not so fine with ISO BMFF encapsulated MP4.
Cyril: MSE says that something
that is not audio or video will fail. The way it works is
that
... dash.js extracts the audio and video parts before feeding
them to the MSE part.
Mike: It applies to any text track of course.
Pierre: A very quick announcement
too - I'm planning to release IMSC.js 1.0.0 by the end of
... this week so if anyone has any comments or input now would
be a great time to raise it.
Nigel: Thanks everyone! See you same time next week. [adjourns meeting]