See also: IRC log
nigel: Today I had planned for an
extra 30 minutes for Dave Singer to chair the agenda item
... on WebVTT but he tells me the preparation isn't ready yet
for that. So we can get the 30 minutes back or use it.
group: Need to finish after 60 minutes
<scribe> scribe: nigel
nigel: Agenda for today is TPAC, TTML. AOB?
group: No AOB.
tmichel: I sent a list today
about a request from Amazon for TTML2 tests, so if we do
have
... up to date tests a pointer would be useful.
gadams: I followed up on that out
of band and sent them a pointer to the TTT github repo
... where there is some TTML2 test material, not W3C sanctioned
at this point.
tmichel: Thank you Glenn!
<gadams> I received the following from Mr Michel. The TTML2 test suite which will be created by the W3C for the purpose of validating the TTML2 specification is not yet available as a whole. However, there are some initial tests found under [1], primarily at [2] and [3]. This project TTT [1] is a work in progress, and support for TTML2 is currently being added, with scheduled completion of March 2017. Note that this project is not originated by W3C itself, but from a 3rd par
<gadams> party, namely my company. [1] https://github.com/skynav/ttt [2] https://github.com/skynav/ttt/tree/master/ttt-ttv/src/test/resources/com/skynav/ttv/ttml2 [3] https://github.com/skynav/ttt/tree/master/ttt-ttpe/src/test/resources/com/skynav/ttpe/ttml2
<tmichel> Glenn maybe you could copy me your email response would be helpfull for a response from me.
<tmichel> thanks Glenn
nigel: I will also respond CC Thierry to explain the status of the TTML2 work and when we will move to implementation/test stuff.
mike: They may or may not mean
TTML2. They may have grabbed the latest spec but actually
... be interested in deployment of TTML1, specifically EBU-TT-D
and IMSC1 so it would be
... helpful to educate them on the status of those specs too,
and their test materials.
nigel: +1
gadams: I don't know for a fact but it's possible that they've been talking to Netflix about TTML2 also.
nigel: There's no harm educating, so I'll do that.
nigel: Registration closes on Sep 2!
<tmichel> FYI Amazon is NOT a W3C Member. We tryed to have them in for Digital publishing but they use their own KINDLE format and not EPUB.
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
tmichel: I will meet Yosuke next Tuesday so I'll ping him directly by voice then.
nigel: Okay, thanks.
tmichel: I will tell him it is a priority because TPAC is getting closer now.
nigel: True. On that topic I will
give myself an action to put together a straw man agenda
... that we can look at, adjust as needed before the
meeting.
<scribe> ACTION: nigel Put together a TPAC straw man agenda [recorded in http://www.w3.org/2016/08/18-tt-minutes.html#action01]
<trackbot> Created ACTION-476 - Put together a tpac straw man agenda [on Nigel Megitt - due 2016-08-25].
nigel: Next week I'm on vacation
so if someone wants to step in and chair this meeting
please
... volunteer, otherwise we have 3 meetings before the face to
face at TPAC.
... Please let me know any agenda items especially if they may
be unexpected.
glenn: We should probably coordinate on getting together on Sunday evening if we want to.
nigel: I'd be happy to meet up at say 2000 based on my arrival time.
glenn: Maybe those who are going
to be there could let you know their locations and we
... could triangulate.
nigel: Yes, sure, okay, also we
usually put hotels etc on the wiki meeting page, so I'll set
that
... up as part of action-475
pal: I'll be there.
action-462?
<trackbot> action-462 -- Glenn Adams to Create issue on ttml2 to add "mapping from other versions and profiles of ttml" -- due 2016-04-28 -- OPEN
<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/462
glenn: No progress on that yet.
nigel: [diverts topic to IMSC temporarily] Pierre do we need to talk about the IMSC roadmap PR?
pal: Not really, but I have been
asked to update the wiki with the latest IMSC
implementations
... and plan to do that in the next week or so.
nigel: Is that also for test resources?
pal: For IMSC 1 we duplicated the
implementation report so that we can add new results,
... test vectors etc so that we could update this new page, so
that may be the right place to
... list IMSC 1 implementations, because they will also have
tests, test reports etc. Then I would
... add a link from the TTML and WebVTT implementations page to
the new IMSC implementations page.
... While I'm doing that I will also apply the PR.
nigel: [Back on TTML2 topic]
There have been some new issues also. What's best to discuss
first?
... Let's cover position syntax first: https://github.com/w3c/ttml2/issues/174
gadams: I was creating some test
content for position and backgroundPosition style
attributes
... that are new to TTML2 and I discovered that there was a
missing rule in the grammar for
... position that effectively made it different than what CSS3
uses for position. CSS3 Background
... and Borders defines the CSS3 property position allows you
to reverse the horizontal and
... vertical components of the position specification, but they
allow vertical first and horizontal second.
... That was apparently an oversight when I transcribed the
grammar.
nigel: How do you know which way round the values are, if that's not a stupid question?
gadams: That's a good question,
particularly because the keyword "center" can apply to
both.
... If you use top or bottom first then it's clear that you
mean a vertical offset.
<gadams> <style tts:backgroundPosition="center left"/> <!-- NG --> <!-- TTML does not support this valid CSS3 expression --> <style tts:backgroundPosition="center right"/> <!-- NG --> <!-- TTML does not support this valid CSS3 expression --> <style tts:backgroundPosition="top center"/> <!-- NG --> <!-- TTML does not support this valid CSS3 expression --> <style tts:backgroundPosition="bottom
gadams: I've pasted above the four test cases (not well formatted) that did not quite work.
nigel: In the issue you use
offset-position-h and offset-position-v but in CSS the
interchangeable
... terms are keyword-position-h and keyword-position-v.
<gadams> https://www.w3.org/TR/ttml2/#style-value-position
gadams: For the 2 component value
it was missing the reverse version. This is expanded
... out as opposed to abbreviated compared to the CSS3
spec.
nigel: At least one of the position keywords must be present to resolve reversed values.
gadams: That's right, and they
have to be something other than "center".
... It is probably worth noting under the grammar that two
linked or center values should
... be mapped to offset horizontal followed by vertical.
nigel: Okay, let's move on to
https://github.com/w3c/ttml2/issues/168
... The background here is that it was unclear if implied
anonymous regions should be
... implemented as new regions or modifications to current
regions. Since then Glenn has
... added a note to the issue to suggest that we add a
parameter attribute to direct the
... behaviour, which sounds like a good idea to me.
gadams: I hadn't thought about
the modification option before you raised it Nigel but
I'm
... willing to give it a try both in the specification and the
implementation.
pal: What's the advantage of having both methods?
gadams: The modification approach
allows region styling to be maintained while applying an
... anonymous set of the region origin and extent, for border,
backroundColor etc.
pal: What's important to
understand is whether the adopters of this syntax also intended
to
... apply region styling with their method. For example if I
use tts:extent and tts:position on a
... p then I cannot apply backgroundColor on that and hope that
it applies. It would be good to know
... what they intended, and to dig deeper before concluding
that we need to support both methods.
gadams: I'm assuming based on
experience and having talked to people that their
understanding
... was based on creating a new region.
nigel: I suspect that the users
of this syntax never added the complexity of region
styling
... or even had temporally overlapping content. For those
documents then it won't make
... any difference which approach is used. For more complex
uses then it would be helpful to have the choice.
gadams: One of the best arguments
against the modification approach is the use of two
... temporally overlapping content elements in the same region.
I also know Netflix has a
... definite requirement for the flavour of creating a new
region.
pal: There are tons of tools that
put extent and origin on elements other than region.
We're
... trying to reverse engineer that into TTML2. The other
option that we should consider is
... simply allowing all the region styling on p and just apply
to the anonymous region.
gadams: That doesn't work because
we have a growing number of properties that apply to
... both content and region, like padding, border, bpd and ipd,
and more. They apply
... independently to both content elements and regions. It
would make the spec much more
... complicated if I have to try to explain how to handle those
dual use properties.
nigel: I'm sympathetic to keeping
things simple, and adding a new parameter is adding
... complexity, but in this case I think there's utility in
both approaches so I would propose
... we add it in and then if necessary later put it in the at
risk list for CR.
pal: Another approach is to say
that if a region is not specified then it would mean 'new'
or
... if a region is also specified then it would mean
'modified'. That would prevent the addition
... of features.
nigel: Ok, we're out of time so let's adjourn for today.
mike: By the way, I've just
submitted https://github.com/w3c/tt-profile-registry/pull/25
... Do we want to review and merge it now?
nigel: Let's not rush for the
sake of the meeting, and look at it offline - if we can add
our
... "LGTM"s etc then we can merge on that basis.
mike: OK, thanks.
nigel: Shall we go ahead with a volunteer chair for next week?
group: Skip next week.
nigel: Okay, see you in 2 weeks
then. Thanks all!
... [adjourns meeting]