W3C

Timed Text Working Group Teleconference

25 Apr 2019

Agenda

See also: IRC log

Attendees

Present
Nigel, Cyril, Gary, Andreas, Pierre, Nigel, Glenn, Thierry, plh
Regrets
Chair
Nigel
Scribe
Cyril, nigel

Contents


<cyril> scribe: Cyril

this meeting

nigel: I've put the charter but not sure there is anything to discuss
... WebVTT IR is important of course
... 2 issues marked for Agenda on TTML2
... and one issue on the profile registry
... I've just added an AOB about an Accessibility Event in June, it's a notification
... anything AOB missing?

TTWG Charter

nigel: the status is that we'll ask to be sent for AC review end of next week
... by that point we should have explainers

thierry: was it presented to w3m?
... before going to AC it should be validate by w3m

nigel: the issues was apply the correct list of chairs, something about webvtt
... it's my expectation that it will need some validation first
... any comment on the current draft charter

atai2: regarding the explainers?
... I haven't been at the meeting where explainers were discussed
... I updated the requirements for 360 subs
... my conclusion is that I don't expect that we'll publish something this year
... discussions are needed to scope this area and write requirements
... I will not be able to write explainers soon

nigel: as the leader of this activity, if you say we'll need to defer this, I'm fine

atai2: the charter is not only covering 2019, up to 2021

nigel: yes but the requirements agreed were for this year

atai2: I just wanted to know if it is ok to delay the explainer until we have better scope

nigel: we have received an incoming liaison from MPEG
... if we defer will there be an impact on these industry timelines
... will we miss the 'right' time to publish something in TTML?

atai2: I don't think we'll be able to deliver something final for this activity in 2019

plh: I forwarded an email on monday
... and I will forward another one

glenn: I reviewed the document this morning
... and it's relevant
... for timed text

nigel: thanks for looking at that

glenn: it basically provides an out of band mechanism for associating a timed text region with a 360 region

atai2: MPEG's document is member confidential

plh: there are 2 documents: the PDF liaison
... and the real meat is the 2 technical documents referenced
... and they are ISO confidential
... as long as we don't copy these documents here we're fine

nigel: I agree minutes are fine

cyril: discussing documents, minuting this discussion, is fine

nigel: andreas do you think it's still best to defer that piece of work, in this group, until 2020

atai2: we should discuss MPEG's document in the group and see what our input can be and see from there
... there is not only ISO/MPEG but the Web Community Group
... I still would like to postpone it
... but I'm also ok to postpone the decision to postpone
... it does not make sense for me right now to produce an explainer

glenn: presumably, this work will be published in a separate module
... so it can have it's own timeline
... and the charter is not explicit about it
... putting end of 2019 as a hard date of all possible rec-track documents is probably not a good idea
... we should focus on TTML2 2nd edition and TTML3
... and the modules that we do finish by the end of the year

ack

nigel: I sent a little summary of explainers to people involved
... anything else on the charter?
... for philippe, at the end of next it will be taken by staff for management review

plh: yes, that will take us 3 to 4 weeks
... it may take less but I cannot guarantee that
... will need horizontal review ...
... I can even circulate it asap if you like to
... it doesn''t have to be 100% final

nigel: you can do that
... I think the remaining issues are PLH to address

WebVTT Implementation Report

nigel: what's the status?

gkatsev: unfortunately, not much time
... we're missing 2 tests: vertical and text combine
... adding should be easy and I should have some time
... also working with Eric to fix display of vertical cues in Safari so that can pass
... text combine, it seems that no browsers support it yet
... not sure what to do about that

<plh> https://github.com/w3c/webvtt/issues

<nigel> plh

plh: there are still some issues not classified in the WebVTT repo
... it would be nice if they are to remain open until the next version
... we should make the decision as a WG to move the spec to PR
... either starting the CfC next week or the week after
... we published the updated CR on April 4th

<nigel> nigel

plh: and unless someone has an issue, we should move ahead

nigel: for me, that is dependent on the IR being ready
... until we have that, we cannot have the CfC

plh: when can we have the IR ready

gkatsev: I already have the spreadsheet
... what would be necessary?

nigel: we need to have good understanding that the test set is complete
... and we create some document describing the set

plh: if we have the spreadsheeet ...

nigel: it might be enough

gkatsev: I can model WebVTT's IR based on the spreadsheet to look like IMSC's IR

<nigel> IMSC 1 IR

plh: that page needs to contain information that cannot go in the spreadsheet
... the criteria is 2 implementations of every feature
... but this is open to interpretation, depending on whether the test is corner case or not
... the tests may not represents an entire feature, if it fails, it may be ok if it is a corner case

gkatsev: I can think of one, right now, font-variant
... CSS as a feature is validated but some CSS properties, white listed, are not supported

nigel: we should actually test that non white listed properties are ignored
... there is a must for that
... and it's important
... we need just one example of a property that is not white listed and see that it is ignored

plh: I agree

gkatsev: font-variant is white listed as part of the font shorthand

plh: is the test testing as part of the shorthand?
... one way to solve this is to check with the CSS WG the status of this
... and see if Chrome considers this as a bug
... or if relying on it is dangerous
... and in the latter case, putting a note in the spec about it
... or if another approach is needed
... gary if you identify other such properties
... let me know
... I can take the action to check with CSS about these properties

gkatsev: I'll take a look at them, and they might be implemented as vendor-prefixed

plh: if it goes in even more details, like units, we might have to go into more details too

nigel: so there is still some work to do to provide the IR, before we can issue the CfC
... anything else?

plh: for the purpose of the charter, I'd like to understand what the group thinks
... it looks like we could publish CR
... and in that case it should be in the charter
... if you are not ok with that, you should speak about it

nigel: for charter, we need to have adequate participation, chair and strategy review

plh: we need to know who is going to maintain?
... and if part of the charter, it needs to be TTWG?

pal: the only question mark in the charter was chairing that
... gary has been active, so we have editing resources
... but the chairing part is not clear

plh: I don't think the chairing part is going to be difficult

pal: the charter has a TBD

plh: and I have an action on that
... either I convince Nigel or I need to convince someone else

gkatsev: I want to see this through so I might step up if needed

plh: reach out to me directly if you want to discuss this

TTML2 and TTML3 Pull Requests

nigel: 2 issues

Add IMSC profile designators to Table 5-2. ttml2#1034

<nigel> github: https://github.com/w3c/ttml2/issues/1034

glenn: I reviewd that and the suggested approach looks reasonable to me and I will prepare a PR

nigel: thank you, done, then

Character-related style properties should not apply to ruby containers. ttml2#1043

<nigel> github: https://github.com/w3c/ttml2/issues/1043

glenn: on this one I posted a PR and haven't got a review

pal: I reviewed it and proposed an alternative PR

glenn: I'll take a look at that

pal: my main question is really: does that match your understanding?
... my alternate PR want to be extremely explicit to see if we agree

glenn: you want to address the possible ambiguity of white space preserve
... in anonymous text spans as children of these containers
... I hate dealing with dueling PR

pal: you raised an interesting point
... another dimension of the issue is XML space
... and applicability, this is not a style property
... my alternative PR does not address that, so maybe we need to merge both PR

nigel: on the point of XML space preserve, a naive interpretation means that there is character content

glenn: basically, we don't want it to be
... it was never the intention that the text content of containers would be treated as text
...glenn: so I tweaked that in my PR
... another thing to keep in mind is that text nodes in containers will be wrapped in anonymous span, so there is no more character content directly in the container

pal: I had not considered that when I filed the ticket
... maybe the 2 PRs are complementary
... nigel, what's your take on the white space issue?

nigel: glenn's PR says that it ignores it
... my comment is broader, it's about any text, not white space

glenn: there are already rules that rule out non-whitespace character
... are you asking about what happens when the content is invalid?

nigel: yes

glenn: in many cases, we don't specify what happens in case of error
... if you are a validating processor, you could reject

nigel: I can see why you have the explicit exception about linear white space

glenn: at that time I had not realized the ambiguity about preserve
... I want to reiterate
... since there are no character content that can appear as a child of container
... my original statement was that we don't need to say anything
... that's when he introduced the ambiguity about white space
... and since that is addressed, I think we don't need to say anything about properties that apply to character content
... that's why I don't think we need to put special case in each property for this

nigel: but they do apply for inheritance

glenn: right
... I pointed that out in the note

pal: but inheritance is not application

glenn: right, application means semantic application

pal: like font-size does not apply to region
... my PR is to make it more explicit
... I do have a question about audio properties

glenn: I do not like the approach of putting special case language in each property

pal: you should consider the amount of change
... there's already special cases

glenn: having repetition of the same statement in many properties is not a good thing

summary: we'll continue reviewing the PR offline to reach consensus

Profile Registry

glenn: we agreed in prior meetings to go ahead and publish
... I need directions on how to do that

nigel: it's published

glenn: excellent

AOB

<inserted> scribe: nigel

Nigel: I just wanted to mention for information that I've been invited to speak about W3C accessibility stuff at an ITU
... event in Geneva on June 7, and plan to attend and do that. There's a link to the event in the agenda - if you want to
... discuss this any more please get in touch with me offline.

Meeting Close

Nigel: Thanks all, we didn't have time to discuss the TTML Profile Registry issue so defer that until next week.
... I'll look at the load for next week and decide 1 or 2 hours, but please let me know if you have additional topics by Tuesday so I can schedule it. [adjourns meeting]

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.154 (CVS log)
$Date: 2019/04/25 16:28:19 $