See also: IRC log
<scribe> scribe: nigel
nigel: Several topics have been
discussed recently: TPAC, IMSC Image profile, TTML2 inline
region semantics
... Any other business or points to prioritise?
pal: Inline region semantics
nigel: You have 1 day left to register if you have not already, and get the advanced rates
action-475?
<trackbot> action-475 -- Nigel Megitt to Contact the chair of the web & tv ig to ask about schedule and joint meeting time. -- due 2016-07-28 -- OPEN
<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/475
nigel: Still no response from Web & TV IG, I have no idea why. Will keep pestering them.
action-476?
<trackbot> action-476 -- Nigel Megitt to Put together a tpac straw man agenda -- due 2016-08-25 -- PENDINGREVIEW
<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/476
https://www.w3.org/wiki/TimedText/tpac2016#TTWG_TPAC_2016
nigel: [runs through draft agenda as it stands]
glenn: I'm going to post a couple
of new issues that we will probably need to discuss.
... One is support for rounded background, and there are a
couple of issues regarding
... more complete support for the more advanced background
attributes. There may be a
... few missing that I think should be added. I'll post some
issues on these to drive agenda items.
... Also on the support for anonymous region generation I've
got some material on that as well.
... It turns out that we need to add back something we used to
have in animation, which
... is the ability to have animation elements point at content
elements, but we changed it
... around to have content elements point at animation
elements. To support anonymous
... inline region for the current region we need an anonymously
generated set element point
... back up to the region.
atai: I can show something on
HTMLCue/VTTCue and we should spend time looking at it.
... There's a question regardless of approach if we should talk
with another WG about this.
... I think something will need to change in the HTML spec
because the spec is format
... independent but the implementation is WebVTT specific, so
to have a format agnostic
... spec for cues then some additions or changes to HTML are
needed.
nigel: Do you know which other groups?
atai: I think we need to know which group is responsible for that part of the spec.
glenn: We could actually draft a
spec or some material in this group, publish it as a Note
and
... throw it over the fence to another group to turn it into a
Rec. That's one option.
atai: What I found out is firstly
that the TextTrackCue should be implemented correctly in
... the browser. They only have VTTCue specific
implementations, though as I read it from the
... HTML spec there should be a generic TextTrackCue
implementation. Initially it may be as
... simple as following the spec as it is written and then add
a property as a payload. Then
... the rest would be done with javascript as a small addition
that would help present any other
... timed content.
nigel: Any other requests for agenda items, or for scheduling?
atai: I think there could be some interest in Web & TV IG in better integration between TTML and HTML5.
nigel: Okay, I'll keep chasing on the joint meeting action.
close action-476
<trackbot> Closed action-476.
nigel: I think we wanted to discuss inline anonymous regions...
glenn: I propose that we adopt a
parameter on a per document basis that defines the
desired
... behaviour for anonymous inline regions. We have already
done half the work, for new
... regions, and need to make progress with current. Then later
if we want to choose only
... one of those options we can do that. That's my
proposal.
... Perhaps we should try to discuss how we develop and
possibly discard features during the spec development process,
at TPAC.
... My general position is that the process step of dropping
features without implementation applies.
nigel: That's certainly the formal position.
pal: I'm not sure if I'm the only one wanting to discuss this.
atai: I'm also interested in this.
pal: This is about how we allocate our limited amount of resources.
<atai> +q
nigel: I will add an agenda topic
for TPAC to consider our deliverable timelines and how
... reliable our estimates are for meeting them, and if we need
to take any action.
glenn: I have scheduled a lot of time to get a CR candidate ready for Nov 1.
atai: I think we should discuss
the balance between adding new features and the challenge
... of implementation when the spec is bigger, from a market
perspective. It is important for
... us to have our specs implemented correctly. We should
discuss our current experiences
... and if this means anything for our spec efforts in
general.
nigel: Summarising, I'll add an
agenda topic for TPAC, and on the subject of anonymous
... regions, Glenn will go ahead with speccing the anonymous
set approach as he proposed.
glenn: Just to note that I
finally merged those PRs.
... One point: when you do propose a PR please don't expand the
RCS keywords otherwise
... I have to manually remove them.
nigel: It could be that there's a problem with the check-in hook then.
glenn: It did not unexpand them
when you checked it in. No problem though, I fixed it.
... Regarding the other PR that I merged I'll post another
issue for clearing up the defined term
... references.
nigel: Thanks, that was for my suggestion prior to merging the PR, that's fine.
glenn: I'm going to try to post a
number of new PRs between now and Lisbon so there may
... be some things for people to start reviewing.
nigel: That would be helpful, thank you.
nigel: The main topic discussed recently is embedded images.
pal: It would be good to have a
way to generate the ISOBMFF files with embedded images,
... which gpac doesn't do today.
glenn: You mean embedded in the BMF not the TTML, right?
pal: Correct. The fact that gpac doesn't support it is not ideal.
mike: It's not just an image
issue. There's currently no tool I've found that can BMF
wrap
... any TTML.
nigel: I thought gpac can do that
- it certainly can do EBU-TT-D though I'm not sure about
... other profiles of TTML.
... Is there any spec implication for IMSC from that
conversation?
pal: I didn't see any.
mike: I have a question: there
are lots of provisions in IMSC 1 that say the content shall
not
... contain X. I'm wondering whether strategically that is a
good thing to do in other profiles.
... Architecturally it's interesting that there are a handful
of forbidden things but nothing
... is forbidden in other namespaces.
glenn: I think it's consistent
with what EBU-TT-D did which seems to be extremely
restrictive
... on what content can be present, presumably to simplify
clients. It's an interesting point.
... When we added support for IMSC 1 to the TTV tool mostly
what we added was checks for
... restrictions, so as far as I'm aware IMSC 1's restrictions
are fully checked by TTV right now.
mike: There's no question that a
validator should check the spec. I'm not trying to revisit
IMSC1.
... Some people want implementations that support things not in
the spec, so that's fine.
... But you can add another namespace and that's fine.
glenn: Because IMSC 1 also says
that any error handling is implementation dependent then
... the normative requirement to prohibit things is very weak.
An implementation may choose
... to ignore them or not. It's useful to note that the core
conformance in TTML says "ignore
... everything you don't understand" basically.
atai: It's an interesting design
approach discussion for profiles. The idea for EBU-TT-D
is
... only things that can be decoded should go in the document.
Other things added may
... change the complete behaviour. That's one approach and that
gives the decoder a guarantee
... that if it's a conformant file then there's only
understandable content that can be decoded.
mike: It's consistent at least in
EBU-TT-D but I'm throwing it out there that there's an
... inconsistency in general in TTML profiles.
atai: There's a question what an
IMSC decoder should do when unexpected vocabulary is
found.
... As Glenn says one option is to say "ignore" and "don't use"
foreign content.
glenn: I said "vocabulary that
the implementation doesn't understand" not "that the spec does
not permit".
... So some implementations can have specific additions.
atai: This is also important to
get some experience back from implementations on which
... design decisions work and which don't work.
mike: What's the status of the media registration now that the note has been updated?
nigel: I don't know. I've not
heard anything. I'm happy to prompt Philippe to remind that
we
... consider the document ready for the next stage of
registration.
... It doesn't look like anything has changed at IANA since Feb
2014.
... I'll remind Philippe. Thanks for merging those PRs.
mike: There are a couple of non-media-registration related issues, which I think were yours Nigel, which we should discuss at some point.
nigel: I'll revisit those and see
if I can make any proposals.
... OK we're out of time, thanks everyone. [adjourns
meeting]