W3C home > Mailing lists > Public > public-tt@w3.org > August 2014

{minutes} TTWG Meeting 14/8/2014

From: Nigel Megitt <nigel.megitt@bbc.co.uk>
Date: Thu, 14 Aug 2014 16:14:18 +0000
To: Timed Text Working Group <public-tt@w3.org>
Message-ID: <D0129CDE.FE38%nigel.megitt@bbc.co.uk>
Thanks all for attending today’s meeting. HTML format minutes can be found at http://www.w3.org/2014/08/14-tt-minutes.html

Text format minutes:


      [1] http://www.w3.org/

                               - DRAFT -

                Timed Text Working Group Teleconference

14 Aug 2014

   See also: [2]IRC log

      [2] http://www.w3.org/2014/08/14-tt-irc


          glenn, jdsmith, mike, nigel, pal, courtney





     * [3]Topics
         1. [4]Agenda
         2. [5]F2F meetings
         3. [6]Timeline for TTML2 publications
         4. [7]MPEG liaison re MIME Codecs parameter
         5. [8]Action items
         6. [9]Issues
     * [10]Summary of Action Items

   <trackbot> Date: 14 August 2014


   <scribe> scribeNick: nigel

   nigel: No AOBs noted

   glenn: regrets for 11th September

   mike: same here

   pal: I'm available on 11th September, so we may be able to do
   some work on that date

   nigel: I'll leave 11th Sep in the diary for now, target 1 hour.

   RESOLUTION: Cancel 18th September phone meeting

   nigel: We'll leave 25th September in the diary, and mark it as
   a 2 hour meeting.

F2F meetings

   nigel: Geneva meeting wiki
   ... I've suggested a 1000 start on the 16th and a 1700 end on
   the 17th - happy to take requests for modifications to this.

     [11] https://www.w3.org/wiki/TimedText/geneva2014#Attendees

   pal: On the Wednesday afternoon there are simultaneous SMPTE
   meetings so I'd rather not schedule topics that I need to be
   involved in then.

   glenn: I propose a 0900 start on the Wednesday.

   pal: I'll arrive on the Monday night.

   glenn: I'm hoping to arrive on Monday night but my travel plans
   are complex!

   nigel: By the way the TPAC registration form has now been

Timeline for TTML2 publications

   nigel: Proposed timeline:
   ... * Week commencing 22nd September: publish TTML2 editor’s
   draft as a FPWD.
   ... * Wednesday 8th October: ‘last issues' closure date for
   TTML2 – any new feature issues beyond this point to be
   addressed after LC or in v.next.
   ... * Thursday 20th November: all issues on TTML2 to be closed
   or deliberately postponed (to v.next or LC review stage). This
   is 3 weeks after TPAC, to allow time for any resolutions made
   in TPAC to be implemented and reviewed.
   ... * Wednesday 17th December: Last Call doc ready to request
   for publication.

   pal: There are a number of CPs assigned to me and my plan is to
   get to them as soon as we get to LC on IMSC, so assuming we do
   that in October I'll be able to start working on the TTML2
   issues that are assigned to me.

   glenn: I may have already tackled some of the issues associated
   with those CPs before then, so my plan for getting to 22nd Sep
   FPWD is to try to crank through all of the open issues
   regardless of the CP backing them.

   RESOLUTION: We will adopt the proposed timeline as outlined

   <scribe> ACTION: nigel draft liaisons to relevant organisations
   for IMSC 1 timeline [recorded in

   <trackbot> Created ACTION-319 - Draft liaisons to relevant
   organisations for imsc 1 timeline [on Nigel Megitt - due

MPEG liaison re MIME Codecs parameter

   <glenn> Issue-305?

   <trackbot> Issue-305 -- Registry of TTML profiles and short
   names -- pending review


     [13] http://www.w3.org/AudioVideo/TT/tracker/issues/305

   glenn: All the action items against Issue-305 were closed a
   draft registry was published that we can fine-tune. It's in the


     [14] https://www.w3.org/wiki/TTML/CodecsRegistry

   mike: Is the intent that it's a 1:1 mapping from the identifier
   to the designator?

   glenn: There's a proposed syntax with a grammar that allows
   combinations of ANDs and ORs to be expressed.
   ... I haven't checked this proposal again after I published the
   TTML2 profile features.

   mike: So these would be content profile designators or
   processor designators?
   ... The note says explicitly 'processor profile'.

   glenn: The TTML2 approach allows a processor profile to be
   inferred from a content profile but not the other way around.

   mike: So for pre-TTML2 profiles?

   glenn: TTML1 documents always declare a processor profile.

   mike: So can you tell the difference in the codecs signalling?

   glenn: No, that codec signalling there, as defined, is only
   signalling processor profile. That raises an issue if its
   correct that a processor profile differs from what can be
   inferred from the document's content profile.
   ... I didn't address that in the registry definition.

   mike: I'll look at that in more detail.
   ... Don't we need a process for updating the table. I think we
   said we'd take it to the group for consensus, but we didn't
   discuss deleting or modifying the table.

   glenn: There is some wording for deleting entries.

   mike: We also need a process for adding entries.

   glenn: The wiki is open to anyone with the right privileges,
   which could be anyone with a W3C account.

   mike: I'm happy to add some verbiage that you think is

   glenn: We can do this in parallel.

   mike: Although the action to respond to SMPTE was closed it was
   done hastily so we should reopen that action or make a new one.

   <glenn> ACTION: glenn to review
   [15]https://www.w3.org/wiki/TTML/CodecsRegistry w.r.t. recent
   ttml2 changes in profile definition mechanisms [recorded in

     [15] https://www.w3.org/wiki/TTML/CodecsRegistry

   <trackbot> Created ACTION-320 - Review
   [17]https://www.w3.org/wiki/ttml/codecsregistry w.r.t. recent
   ttml2 changes in profile definition mechanisms [on Glenn Adams
   - due 2014-08-21].

     [17] https://www.w3.org/wiki/ttml/codecsregistry


   <trackbot> action-283 -- Nigel Megitt to And dsinger to respond
   to mpeg liaison -- due 2014-05-01 -- CLOSED


     [18] http://www.w3.org/AudioVideo/TT/tracker/actions/283

   <glenn> ACTION: mike to update
   [19]https://www.w3.org/wiki/TTML/CodecsRegistry to add proposed
   language about process for changing registry [recorded in

     [19] https://www.w3.org/wiki/TTML/CodecsRegistry

   <trackbot> Created ACTION-321 - Update
   [21]https://www.w3.org/wiki/ttml/codecsregistry to add proposed
   language about process for changing registry [on Mike Dolan -
   due 2014-08-21].

     [21] https://www.w3.org/wiki/ttml/codecsregistry


     [22] http://lists.w3.org/Archives/Public/public-tt/2014Apr/0025.html

   mike: After Glenn and I have worked on this we should follow up
   with a link to the wiki.
   ... The liaison was sent after the April meeting and it didn't
   appear in the June meeting, so I'll check if anyone saw it, and
   we can take appropriate action depending.


   <trackbot> action-290 -- Mike Dolan to Describe the history for
   how we got to the signalling needed for mpeg that came from
   dece for issue-305. -- due 2014-05-29 -- OPEN


     [23] http://www.w3.org/AudioVideo/TT/tracker/actions/290

   close action-290

   <trackbot> Closed action-290.

   nigel: Do we have clear requirements for this request?

   mike: I'm not sure - for example the ability to restrict or
   extend specific features.

   glenn: Feature restrictions and extensions are in TTML2.
   ... There are also new top level parameters to express the
   policy for processors to widen or narrow the semantics.
   ... For example if you define a color feature extension that is
   a restriction, but the processor already supports the
   unrestricted version, it could widen the semantics of the new
   feature or extension to claim support for the extension.

   mike: Okay I'll take a look at this.

   glenn: Issue-263 is relevant here too and we were waiting for
   mike's review on that.
   ... That's also pending review.


   <trackbot> issue-263 -- profile feature set may not match
   intended feature constraints -- pending review


     [24] http://www.w3.org/AudioVideo/TT/tracker/issues/263

   mike: As long as this is now defined then I'm happy with this.

   glenn: Do we want to go back and change SDP-US to make use of
   any of these new mechanisms e.g. to clarify that its definition
   of the color requirements is a restriction on the existing
   color feature.

   mike: Probably everything using profiles now needs to be
   revisited, and SDP-US is one example, as is color within that.

   nigel: I propose that we open a new issue to update SDP-US when
   the dust has settled on TTML2.

   glenn: SDP-US is only a Note so we would either have to move it
   to Rec track or issue a new note. That may impact on the

   nigel: We're not sanctioned to move it to Rec.
   ... But updating the note on our SDP-US is in the charter so we
   can go ahead and do it.

   mike: I'm not sure we should spend time on SDP-US when the
   proponent is no longer active.
   ... That was Sean.

   jdsmith: I'm pondering how we keep this up to date. The -US tag
   indicates alignment with FCC requirements.

   mike: Plus there were some provisions that were peculiar to the

   <scribe> ACTION: jdsmith Indicate preference for updating
   SDP-US for TTML2 [recorded in

   <trackbot> Created ACTION-322 - Indicate preference for
   updating sdp-us for ttml2 [on Jerry Smith - due 2014-08-21].

   mike: The Issue-305 MPEG liaison did get lost - sending to the
   secretary often doesn't work. Let me know when we're ready to
   send an update.

   reopen action-283

   <trackbot> Re-opened action-283.

   action-283: This needs to be redrafted and sent via mike, to
   reflect the draft registry page and updated profile mechanisms.

   <trackbot> Notes added to action-283 And dsinger to respond to
   mpeg liaison.

Action items


   <trackbot> action-295 -- David Singer to Socialise the geneva
   f2f meeting amongst the webvtt community -- due 2014-06-05 --


     [26] http://www.w3.org/AudioVideo/TT/tracker/actions/295

   close action-295

   <trackbot> Closed action-295.


     [27] https://www.w3.org/2002/09/wbs/34314/TTWG_Geneva_F2F_Planning/results

   Courtney: I think we've done all we can do to publicise the
   meeting in the WebVTT community. Andreas and I can represent
   the WebVTT side of things for that discussion.


   <trackbot> action-309 -- Nigel Megitt to Propose a timeline for
   ttml 2 fpwd -- due 2014-08-14 -- PENDINGREVIEW


     [28] http://www.w3.org/AudioVideo/TT/tracker/actions/309

   close action-309

   <trackbot> Closed action-309.


   <trackbot> action-313 -- Glenn Adams to Create erratum for
   issue-314 -- due 2014-08-07 -- PENDINGREVIEW


     [29] http://www.w3.org/AudioVideo/TT/tracker/actions/313


   <trackbot> issue-314 -- Temporally active is not defined for
   regions -- closed


     [30] http://www.w3.org/AudioVideo/TT/tracker/issues/314

   close action-313

   <trackbot> Closed action-313.


   <trackbot> action-290 -- Mike Dolan to Describe the history for
   how we got to the signalling needed for mpeg that came from
   dece for issue-305. -- due 2014-05-29 -- CLOSED


     [31] http://www.w3.org/AudioVideo/TT/tracker/actions/290


   <trackbot> action-307 -- Mike Dolan to Assign priorities to
   change proposals owned by mdolan -- due 2014-07-24 -- OPEN


     [32] http://www.w3.org/AudioVideo/TT/tracker/actions/307


     [33] https://www.w3.org/wiki/TTML/changeProposal024


   <trackbot> issue-167 -- allowing attributes that have no
   semantics on specific elements -- open


     [34] http://www.w3.org/AudioVideo/TT/tracker/issues/167

   glenn: I can handle issue-167, so maybe the fix is to change
   the owner of the CP to me.

   mike: Are you going to codify a proposal?

   glenn: I'll just do it and then we can review.

   mike: What is your intent here?

   glenn: It's to add a single sentence paragraph to the header of
   the style processing semantics explaining that even though
   attribtues can be present on elements to which they do not
   ... for inheritance purposes, then if they are not inherited
   they are effectively ignored. I'm not sure about prohibiting
   them from being present.

   mike: My issue was to prevent a mass of attributes on the tt:p
   elements for which they have no meaning and can not be used by
   child elements.
   ... I'm leaning towards forbidding them.

   glenn: It's something that could be readily tested in the TTV
   product but could not be described in the schema so I don't
   really have a huge problem with it. We can also make it a
   ... and they're handled in TTV by issuing warnings as opposed
   to errors.
   ... Also one of the attributes is origin, and we have now
   defined the meaning of origin and extent on p and div.

   mike: At the time that was of no use but people did it and
   invented semantics for it. That example no longer applies.

   glenn: The question is if we make a SHALL NOT or a SHOULD NOT.

   mike: I'd be happy with a SHOULD NOT. By the way the schema
   could be updated for this, though it would be a significant
   change. I was concerned about attributes that no child can make
   use of - it should be just wrong.

   glenn: For example zIndex and opacity - we've restricted those
   to regions only, even though opacity would have a CSS mapping.

   mike: Maybe they don't all have the same rule. The minimum bar
   would be SHOULD NOT and on a case by case basis they could be

   glenn: Since the semantics are to ignore them there would be no
   difference in processing between SHOULD NOT and SHALL NOT. The
   only difference would be, for a validating processor,
   ... if we expressed it as SHALL NOT then a validating processor
   would need to check and potentially reject the document. There
   would be at most a warning if we make them SHOULD NOTs.

   mike: The more mature it is and the more validating tools are
   out there the less this is a practical concern.

   <glenn> [35]https://github.com/skynav/ttv/issues/12

     [35] https://github.com/skynav/ttv/issues/12

   mike: Let's go with a sentence providing a warning to authors
   that there are attributes that have no semantic meaning and are
   potentially not even inheritable.

   glenn: I propose to go with SHOULD NOTs and open an issue on

   mike: I'm okay with that - anything we can do in the spec to
   make clear that syntactic permission doesn't imply processing

   <glenn> see [36]https://github.com/skynav/ttv/issues/12 for TTV
   issue (already registered)

     [36] https://github.com/skynav/ttv/issues/12

   nigel: Does this need to be pre-FPWD or pre-LC?

   glenn: I'll do this pre-FPWD.

   nigel: I'll make this a priority 1 then.

   action-307: This is a priority 1, and reassigned to glenn

   <trackbot> Notes added to action-307 Assign priorities to
   change proposals owned by mdolan.

   close action-307

   <trackbot> Closed action-307.


   <trackbot> action-311 -- Nigel Megitt to Check with mdolan on
   status of issue-263 and issue-305 etc -- due 2014-08-07 -- OPEN


     [37] http://www.w3.org/AudioVideo/TT/tracker/actions/311

   close action-311

   <trackbot> Closed action-311.

   nigel: adjourning until xx:12
   ... And that's now!



   <trackbot> issue-263 -- profile feature set may not match
   intended feature constraints -- pending review


     [38] http://www.w3.org/AudioVideo/TT/tracker/issues/263

   close issue-263

   <trackbot> Closed issue-263.


   <trackbot> issue-311 -- Note on progressivelyDecodable is not a
   sentence -- pending review


     [39] http://www.w3.org/AudioVideo/TT/tracker/issues/311

   mike: on issue-263 we didn't actually do anything on SDP-US.

   glenn: We addressed the requirements for TTML2 but nothing on
   SDP-US. I'd suggest that this is a new issue, rather than
   keeping the old one open.
   ... For example: "Update SDP-US to make use of new mechanisms
   to define..."

   <scribe> ACTION: nigel to update issue-263 to target product
   TTML2 and open a new one on SDP-US. [recorded in

   <trackbot> Created ACTION-323 - Update issue-263 to target
   product ttml2 and open a new one on sdp-us. [on Nigel Megitt -
   due 2014-08-21].


   <trackbot> issue-311 -- Note on progressivelyDecodable is not a
   sentence -- pending review


     [41] http://www.w3.org/AudioVideo/TT/tracker/issues/311

   close issue-311

   <trackbot> Closed issue-311.


   <trackbot> issue-330 -- Is Presented Region a synonym for
   temporally active region? -- pending review


     [42] http://www.w3.org/AudioVideo/TT/tracker/issues/330

   close issue-330

   <trackbot> Closed issue-330.


   <trackbot> issue-334 -- Misuse of style property
   characteristics with ttp:progressivelyDecodable -- pending


     [43] http://www.w3.org/AudioVideo/TT/tracker/issues/334

   glenn: I'm happy with this as long as it doesn't present the
   parameter with the appearance of a style attribute.

   pal: I reverted to the original prose.

   glenn: If you look at the current TTML parameter definition we
   use a particular form of language, which would be good to

   pal: I'll make a change accordingly.

   reopen issue-334

   <trackbot> Re-opened issue-334.

   <glenn> issue-334: change language to "If
   ittp:progressivelyDecoable is not specified, then it must be
   considered to be equal to ..."

   <trackbot> Notes added to issue-334 Misuse of style property
   characteristics with ttp:progressivelyDecodable.


   <trackbot> issue-288 -- Rules for splitting and accumulating
   documents -- open


     [44] http://www.w3.org/AudioVideo/TT/tracker/issues/288


     [45] https://www.w3.org/wiki/TTML/changeProposal025

   nigel: I have proposed a set of rules and an algorithm for
   combining documents.

   glenn: I'd suggest this as ttm:group to be significant only on
   the tt:tt element.
   ... It may be more effective to define this in terms of an ISD
   sequence, but I can see applicability beyond that.

   nigel: I agree that restricting it to ISD sequences is too

   glenn: There's a question whether this is a parameter or a
   metadata item. Parameters are used to direct the processor, and
   I'm not sure this group mechanism would have processor
   ... Unless we define processor semantics that are mandatory to
   use in certain circumstances I would make it a metadata
   attribute because it feels more like that to me.

   nigel: It's a specific processor parameter for a 'combiner'.

   glenn: That suggests to me that it's not a general parameter
   but a more specific one, which tells me it's a metadata item
   not a parameter.

   <mike> As expressed on the reflector, I remain of the opinion
   that this feature is not needed. That said, the proposal does
   no harm provided its semantics are clear.

   <mike> I have to leave for another meeting

   glenn: I've been trying to restrict parameters in TTML to
   things that area applicable to all processing and
   transformation not a specific set of transformation scenarios.

   nigel: I was at pains to make this not affect processing of any
   single document.

   glenn: In that case it would be metadata e.g. in the ttm:

   nigel: I will update the CP, and make it an NCNAME.


   <trackbot> issue-307 -- Conformance language and processor
   profile rather than content profile. -- open


     [46] http://www.w3.org/AudioVideo/TT/tracker/issues/307

   nigel: I've raised a new CP, CP28 whose main outcome is to
   create a new text profile that does not include the complexity

   pal: The summary is one thing, but the CP has lots of other
   statements and proposals that go beyond.
   ... We could just amend the CP to reflect that end result. If
   that's the desire then it should be captured as such.

   glenn: I've renamed the CP to include "IMSC" in the name.

   pal: I think we shouldn't jump ahead to the changes needed to
   achieve the goal.

   nigel: I was attempting to be helpful to the editor in
   proposing the edits needed.
   ... Are you objecting to feature designators for the HRM?

   pal: I don't think its necessary: another way to achieve the
   goal is to factor out the HRM section and add it into 4.2 and
   4.3, and add 4.4 for Unconstrained Text Profile.

   nigel: Okay, that might work.
   ... If we want to apply different edits we can change the Edits
   to Apply section.

   pal: Also the CP references issues that haven't yet been filed.

   nigel: Agreed, we can take no action on issues that have not
   been filed.

   pal: For example ebutts:linePadding has not been filed as an
   issue, so we should consider this if it is filed.

   nigel: I'm happy to put this on pause until the issues for
   which it would be of use are raised.

   pal: I can make some edits on this CP.
   ... The additional constraints in some distribution mechanisms
   are orthogonal to the HRM.

   nigel: The HRM may block resolution of future issues, but we
   can wait until those are filed.

   pal: You can make highly complex but low data size documents.

   nigel: Questions if this is really true.

   pal: You can with animations.

   glenn and pal: discuss the generation of ISDs as they may be
   created for animations.

   nigel: Happy for pal to make an alternate proposal that we can


   <trackbot> issue-310 -- Forward reference rule doesn't take
   into account child elements -- open


     [47] http://www.w3.org/AudioVideo/TT/tracker/issues/310

   glenn: In TTML1 and TTML2 there is no forward referencing
   defined other than implicitly via the expression of parallel

   pal: Can style references be made forwards?

   glenn: Content elements can not reference forward style
   elements however inside of a styling element itself chained
   referential styling can have forwards references, but not
   circular ones.
   ... That's the only case - we can simply make that require
   backward references only.

   pal: That's what §5.5.2 point 4 is specifically meant to deal

   glenn: Is that really related to progressive decoding of
   content though, because you have to have decoded the <head>
   before you get to the content.

   pal: I was told that some implementations would not like
   forward reference styles. Since it didn't seem a huge burden
   and in some ways a good thing it doesn't seem harmful.

   glenn: To call it out specifically, since there are only two
   example of potentially non-progressively decodable that we've
   found: chained styles and timings we could call them out
   explicitly rather than taking the generic approach.
   ... That would make it a little more concrete for authors and
   implementors to understand.

   pal: This is to state points 2 and 4 more concretely?

   glenn: Either that or add a note to call out the two cases
   where those constraints might be validated.

   pal: If you can jot that note down it would be a great

   <scribe> ACTION: glenn draft a note for IMSC 1
   progressivelyDecodable to make concrete what authors should
   take into account [recorded in

   <trackbot> Created ACTION-324 - Draft a note for imsc 1
   progressivelydecodable to make concrete what authors should
   take into account [on Glenn Adams - due 2014-08-21].

   glenn: there's still an outstanding point about the meaning of

   action-324: Also draft wording around which 'documents' are
   being referred to in the numbered list.

   <trackbot> Notes added to action-324 Draft a note for imsc 1
   progressivelydecodable to make concrete what authors should
   take into account.

   nigel: adjourns meeting

Summary of Action Items

   [NEW] ACTION: glenn draft a note for IMSC 1
   progressivelyDecodable to make concrete what authors should
   take into account [recorded in
   [NEW] ACTION: glenn to review
   https://www.w3.org/wiki/TTML/CodecsRegistry w.r.t. recent ttml2
   changes in profile definition mechanisms [recorded in
   [NEW] ACTION: jdsmith Indicate preference for updating SDP-US
   for TTML2 [recorded in
   [NEW] ACTION: mike to update
   https://www.w3.org/wiki/TTML/CodecsRegistry to add proposed
   language about process for changing registry [recorded in
   [NEW] ACTION: nigel draft liaisons to relevant organisations
   for IMSC 1 timeline [recorded in
   [NEW] ACTION: nigel to update issue-263 to target product TTML2
   and open a new one on SDP-US. [recorded in

   [End of minutes]

    Minutes formatted by David Booth's [55]scribe.perl version
    1.138 ([56]CVS log)
    $Date: 2014-08-14 16:10:39 $

     [55] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm
     [56] http://dev.w3.org/cvsweb/2002/scribe/
Received on Thursday, 14 August 2014 16:14:51 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 5 October 2017 18:24:17 UTC