See also: IRC log
<nigel> scribe: nigel
Nigel: Today we have one issue
marked for agenda for TTML Profile Registry.
... TTWG Future Reqs if there's anything to cover,
... F2F meetings next week
... TTML in RTP IETF submission
... Is there any other business, or any specific points to
raise?
... I didn't have reason to add WebVTT to the agenda before,
but there's been some discussion on
... the reflector today, so we can add it if that would
help?
Gary: Yes, we can cover WebVTT please.
Nigel: OK
group: [nothing more for the agenda]
github: https://github.com/w3c/tt-profile-registry/pull/58
<cyril> scribe: cyril
nigel: PR to fix etx2
... I raised it, andreas reviewed it, open for 13days
... it's ready for the editors to merge
<nigel> SUMMARY: Editors to merge tomorrow at the earliest assuming no new objections
nigel: Is there anything specific
to discuss?
... there has been some discussion on embedded images
... it's fine offline
... no comment on that point
nigel: I have updated the
agenda
... we need some time on WebVTT
... I am not sure what the best time would be
gkatsev: day 1 would be good
cyril: If the discussions on the
requirements I raised could be discussed at a convenient time
for me
... like 4pm geneva time
nigel: day 1 is mainly those
requirements, TTML2, modularization ...
... live demo as part of the PTS event
... and the WebVTT implementation report
... day 2 in the morning is the joint meeting about live
ttml
... and how we handle EBU defining semantics
... like multiRowAlign but other contributions
... we need to think about what the options are
... if you have any other topic let me know
nigel: we had a bit of feedback
from Mike
... my colleague James just updated a new draft
<nigel> RTP Payload for TTML Timed Text
nigel: one of the thing that
generated discussion is the media type definition
... there seems to be a best practice IETF document saying that
you should include the media type registration for the payload
format
... that explains why the draft copies it
... that's a bit bizarre (to me, personally)
... but it seems a recommended practice
... the change control now says W3C has change control over
this specification [i.e. the media type registration]
... there has been a review/comments from members of the IETF
group as well
... it's a good sign, has momentum
nigel: gary and thierry have been discussing this
gkatsev: the status is that
unless there is progress on the spec, it will be removed from
the charter and not added back soon
... my question is what would constitute progress
tmichel: the question is not
really about the progress on the spec, because it's at CR and
stable
... the issue is really about the implementation report
... I've looked at wpt.fyi
... I agree that for APIs, most of the tests pass 80%
... but the rendering is at about 1% or 2% close to zero
... it needs to be run manually
... I discussed with Philippe
... we need a proper implementation report
... showing things are implemented twice
... if we don't progress to PR, it's unlikely that it will stay
in the charter
... if the implementation report progresses, the transition to
PR should be ok
gkatsev: I'm actually working on going through the rendering tests across browsers
tmichel: I've done a bit
too
... my conclusion is that Chrome it was good, Edge was bad and
FF was so and so
... I didn't test Safari
gkatsev: Safari is pretty good as well
tmichel: we need to know where we are
gkatsev: that is what I'm working on and will have results next week
tmichel: you have to know the past and that the efforts have been slowing down, so thank you for joining the group and putting efforts
gkatsev: I totally
understand
... I would like to see it through if we can
nigel: one question I have is:
how would you establish what counts as a feature?
... as far as I know, WebVTT does not have a list of features
in the spec
... TTML has it but a lot of other specs don't
... what's your approach Gary?
gkatsev: the way they set it up
in WPT is that each test is a feature
... the rendering tests have standard cues, or 2 cues
overlapping
... it seems that the tests cover a pretty wide variety of
"features"
... for example has a separate file per unicode characters
nigel: if each test is a feature, how do you assess completeness of the testing vs the spec?
gkatsev: as Silvia said, Apple
funded the development of the tests based on the spec
... I'm assuming it's fairly complete
... but I would need to verify that
... but it seems the big and important parts are covered
nigel: it's fair to say that
there is always a bit of judgment from the group on what
features are and on the coverage
... if you do that review and see areas are weaker, it would be
good to bring that to the group
atai2: could you check if after the tests have been completed, if there were changes to the spec
gkatsev: yes, I can check when
the last change to the spec was and the last change to the
tests was
... I doubt there was substantive change to the spec after the
tests have been done
atai2: that would be good to review it
nigel: anything else on webvtt?
gkatsev: not for now