W3C

Timed Text Working Group Teleconference

21 Feb 2019

Agenda

See also: IRC log

Attendees

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

Contents


<cyril> scribe: Cyril

This meeting

TTML Profile Registry Actions, Pull Requests and Issues

nigel: I've looked at the issues and Glenn does not seem to have update the issues
... there are still 2 PR: his and mine on his
... Mike nothing to add on that

mike: no I kind of lost track of what was going on

nigel: I noticed something that relates to the RTP
... the codecs parameter specifies + and |
... but it references processor profile combination in TTML2
... I've added issue #62
... it seems we can't make progress without glenn

pierre: is anything blocked here?

nigel: yes
... we should have published an update of the profile registry regarding the publications on Nov 8th

pierre: IMSC1.1 is there in the editor draft

nigel: but not in the published one
... we should be really trying to update it to make sure it has the correct set of profiles

pierre: Mike, what do you need to get this published

mike: I am not the editor anymore
... because the document has turned into many issues and not the simple document I thought I would be editing

pierre: people care about the list only
... can we keep it simple

nigel: yes, but glenn does not seem to have this view
... and glenn is the only editor

pierre: we should revisit in a month and possibly change strategy

nigel: straw poll, does anybody else share my view that it is important to publish

pierre: yes, I thought this has been publish

cyril: yes

pierre: if it's a resource isssue, I'm willing to take a stab at it but would simplify the document a lot
... the table that really matters is 4.1
... all the rest of the text is really not useful
... especially if it causes issues

nigel: there is consensus that we should get this done
... I'll mark this item on the agenda in 2 weeks

cyril: I think we should be stronger
... if there is no update in X weeks, we should change strategy and editor

nigel: how many X is?

cyril: I would say X=2 weeks with a possible extension

plh: my philosophy is: if the ED is better than TR, publish a new version no matter open issues

pierre: I agree with cyril

nigel: ok works for me

TTWG Future requirements

nigel: no update this week

<inserted> scribe: plh

Cyril: not clear how we're converging here. We had proposals and ideas, but how do we decide to move forward?

<inserted> scribe: cyril

cyril: not clear to me how we turn reqs into text

nigel: if WD is 1st of June/July, the first thing to do is to write explainers

cyril: I fear that writing explainers will delay the spec

nigel: if you want to go and propose text go ahead

pierre: in general, whoever submitted the requirement is on the hook to provide the first pass text
... until that happens, nothing can happen

cyril: clearer now thanks
... what about modularization
... should the proposed text be in the form of a module

nigel: if you think the text you want to propose fits in a module, then yes propose it as a module
... I'm happy to schedule time in the agenda for early discussion
... if any group steer is heplful

<plh> https://www.w3.org/TR/2006/NOTE-ttaf1-req-20060427/

plh: do you mind if I retire the UC and Req for TTML1
... right now it appears in the page of W3C
... I want to say this document is only here for historical purposes

nigel: TTML specs reference it

plh: if you think it still matters, then I won't change it

nigel: yes we reviewed it

plh: ok, then I won't touch it

TTML in RTP IETF submission

nigel: we discussed it last week
... there has not been any update
... but the ongoing work will fix the current issues
... the new IANA section will say

<plh> https://datatracker.ietf.org/doc/draft-sandford-payload-rtp-ttml/

nigel: W3C consideration sections will disappear
... and IANA section will say "No IANA action" and the duplicate media type registration text has been removed

mike: do you want to at least make a note that the registration details are in the W3C

nigel: yes it is somewhere else in the document

mike: oh I would put it here

nigel: it's in section 7

mike: odd location, but ok

nigel: the other change is around the codecs parameter in the SDP
... we will say "shall be present" and mention processor profiles
... I'm planning to use the profile combination logic of TTML2

plh: what was the incentive to copy the IANA registration

nigel: we will remove it

WebVTT Implementation Report

gkatsev: I've been working on it, submitted 2 PR to WPT
... just approved
... one fixes the preferences for line start and line end
... implementations were correct
... one for the references for white space pre
... the rendering tests go up to 81% completion
... the big thing on which I'm working now is region
... unfortunately, the collision avoidance tests are not passing
... difference of implementation
... Chrome only does it in some cases
... snap-to-line is true
... the tests we have use snap-to-line false
... some browsers like Firefox also implement it but not according to my reading of the spec
... you should choose the one above the cue first
... I can raise issue against browsers but not sure it will be fixed soon
... and that would put the IR at risk

pierre: I'd love to pick your brains on this topic
... it is really a part that I find weird and confusing
... the author specifies a position but not completely
... I think it's a pretty complex algorithm
... but the author should have complete control over position
... collision avoidance should not be in the spec

gkatsev: looking at the test it seems to have been even more complicated

pierre: I'm not surprised that implementations do different things
... one option is to remove it and say collision avoidance is up to the implementation

gkatsev: there is an accessibility issue
... you want to always show the caption if you can
... maybe on snap to line false we could remove that

pierre: I don't dispute the overall goal
... it's ok to put that burden in the implementation rather than in the spec
... I haven't seen cases with multiple tracks
... maybe forced but it's mutually exclusive with subs

gkatsev: I looked at WPT.fyi tests
... it seems that the API passing tests jumped from 80% to 90% and so they are up to fixing things
... one question: assuming WebVTT is removed from the charter
... what are the options for WebVTT going forward?

pierre: have we eliminated just trimming the spec
... to match what implementations do?

nigel: not sure we have time because that means publishing CR

pierre: if that were the strategy, it would help scope the charter

plh: republishing the spec by just listing features at risk, I don't see why it would be long

nigel: just based on experience

plh: it's not like the charter is ready to go in a month

nigel: the draft to review in march

plh: if we want to take the time, we have the time
... but if nothing happens until now and the time the charter goes to AC
... I will oppose mentioning WebVTT in the charter
... and it would simply remain in the community group

pierre: going back to what drove having both WebVTT and TTML in the same group, is because having them in two separate groups would have been worse
... putting browsers in the path of publication is not a good idea because that is out of our control
... is there an appetite and interest to trim the spec to match what is done today

nigel: there is not a single simple answer to that
... it depends what would be removed
... for example things that are needed to meet FCC reqs would be a problem

pierre: what do you think of trimming?

gkatsev: probably possible
... but removing things like region is very bad
... it's needed for FCC compliance and accessibility

pierre: it's even worse to have something that does not match reality?
... for example ISMC1 did not support Japanese, but IMSC1.1 does
... it's better to have a first spec that reflects reality and then have a second version

gkatsev: maybe that's better
... I'm not sure

plh: you can branch the current spec
... and do what pierre is suggesting
... you keep an ED with everything to start a v2

pierre: what is bad is having specs that are meaningless

gkatsev: maybe that's the best goal

plh: I think you should have asap a list of features that could be marked at risk
... for review
... that is assuming that trimming would not take months
... and based on the IR would go to PR quickly
... we can even publish WD for v2 before v1 is published

gkatsev: that sounds like the best course of action

TTWG Charter

plh: the draft now matches the template

nigel: I have raised 3 issues
... 34 to 36
... 34 is to update regarding TT Requirements
... specifically live contributions and audio profile
... 35 is about modularization
... to make it clear that deliverables will be modules
... and to allow the group to decide which module will be on the rec track or not
... 36 is a placeholder for WebVTT
... I think that is the scope of what needs to change in the charter
... let me know if anything else needs to change by raising an issue on the repo

<nigel> Draft TTWG Charter repo

<plh> https://w3c.github.io/charter-timed-text/

nigel: my timetable is to update the charter in the next weeks, except WebVTT

Hosting additional test/example resources

nigel: pierre received some material from Fox
... that would be useful
... not as test but generally useful

pierre: Fox have put 2 test reel, one for IMSC 1 text, IMSC 1 image, IMSC 1.1 text and image
... it's IMSC documents and associated video and render
... this is not a unit test, not covering all variations
... but covers several reasonable variations
... IMSC 1 and IMSC 1.1 unit tests are not good sample materials, missing video and sync
... that test material is really good complement to the IMSC unit tests
... they've offered it to W3C
... it would be more valuable if hosted by W3C
... available under BSD licence
... my recommendation is for W3C to accept the offer and host the content as part of the WG work

plh: my initial reaction is to say no because it's already hosted on GH
... you will need to convince me
... why should we do it for this WG?
... do we need to insure maintenance

pierre: a member of the WG told me their lawyers have cleared using materials from W3C not from general GH
... even if the different site uses BSD
... another reason is for W3C to bring more people to W3C
... because the reel is of good quality
... on the maintenance the folks at Fox intend to maintain it because they use it
... so the work by this group would be minimal

plh: I don't care if Fox maintains it, I need commitment from the group (it may be delegated to Fox)

pierre: I'm happy to commit to that

nigel: if someone is implementing IMSC they need to find this resource
... we also have a wiki page and we could add it there
... I'm not sure what the right place in W3C is
... maybe MDN could also be a good place, for developer guidance

pierre: linking from MDN would be a good thing but this does not solve the license/lawyer issue

cyril: I think this is important to have this test in W3C space because this is testing video + IMSC and it is the first

nigel: a lot of people have done it

pierre: yes but no one has made it available to W3C
... HbbTV has nice tests apparently but they are not available
... I think the wiki is fine

plh: why not create a separate repo

pierre: i'm fine with that too

nigel: we'll try to resolve offline. I don't have objections

plh: I'm happy to have the group adopt those tests
... without the WG support I cannot take it

nigel: does anybody have an objection to W3C hosting it?
... seems not
... then plh you can go ahead

plh: ok I'll follow up with Pierre

nigel: I added another AOB
... you have a poll in your inbox
... as to how we get to meet in september
... please reply

<nigel> Poll

Meeting close

<nigel> scribe: nigel

Nigel: Thanks everyone, apologies for running 10 minutes over. [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: 2019/02/21 17:48:21 $