[minutes] teleconference 20090123

Available at

Text version:

                Timed Text Working Group Teleconference

23 Jan 2009


          Philippe, Glenn, John, Sean, David, Franz





     * [3]Topics
         1. [4]ISSUE-3
         2. [5]ISSUE-5
         3. [6]ISSUE-4
         4. [7]ISSUE-6
         5. [8]ISSUE-7
         6. [9]ISSUE-8
         7. [10]ISSUE-10
         8. [11]ISSUE-11
         9. [12]ISSUE-18
        10. [13]ISSUE-19
        11. [14]ISSUE-20
        12. [15]ISSUE-26
        13. [16]ISSUE-27
        14. [17]Next steps
        15. [18]Namespace style
     * [19]Summary of Action Items


   Sean: we have 2 yes, 2 no
   ... the goal was to required the ttm: namespace elements to be

   Glenn: I believe they should be allowed outside the metadata

   Sean: the reason for this to come up: it's just easier if all the
   metadata is in one place
   ... having it all over the place makes it more complex

   Glenn: the rationale was to support farm metadata as well.

   Sean: because the metadata element is called "metadata", I assumed
   that all metadata will go there

   Glenn: part of the reasons for having the metadata element was to
   support transformation processor to group metadata
   ... ie having some grouping mechanisms
   ... if a transformation combines metadata from two sources into one
   ... it might group them by origin in the result document

   Sean: it's not about having the metadata grouping element
   ... if title is metadata, they should be in the metadata element
   ... if not, then it should be moved out of the metadata namespace

   Glenn: it creates work to make a change now and it doesn't seem a
   significant issue to change now

   Sean: that's a cognitive issue. I'd expect to find all the metadata
   would be in the metadata element
   ... how much work is it to move title out of the metadata namespace
   or within the metadata element

   Glenn: probably not that complex. change in the schema. that's
   substantive change, but it wouldn't be that difficult

   David: I don't have a strong feeling one way or the other
   ... if we're going to put title in ttm, we identified it as metadata
   already, why moving it into metadata?

   Sean: since no one feels strongly, let's leave as-is


   Sean: ISSUE-3 was about more than one metadata element, and ISSUE-5
   is about direct children

   Resolution: ISSUE-3 and ISSUE-5 closed without change unless Geoff


   Resolution: we should allow the ttm attributes on the region element

   ACTION: Glenn to allow ttm attributes on the region element
   [recorded in


   Sean: the idea here is that we define fixed set of roles and if we
   want to extend, we would need to add values
   ... agent is fairly complicated structure and you can point to that
   ... having a role element would allow the same
   ... there is no user defined role in the spec

   Glenn: there is an extension mechanism.

   Resolution: close ISSUE-6 without changes

   David: peharps we could simplify the actors as well...

   Glenn: are you suggesting a ttm:actor attribute?

   David: yes, ttm:agent is an IDREF that has to refer to the agent

   Sean: should we wait on this and revisit issue-6 later?
   ... let's keep close and reopen it later if necessary


   Sean: we inherited it from SMIL. that's the only abbrev form we have

   Glenn: we have an appendix that describe where our vocabulary comes
   ... ttm:desc comes from svg:desc

   Sean: why from svg?

   Glenn: we took a couple from them: metadata, set, desc, title comes
   from SVG
   ... for the attributes, we referenced SMIL for some of them

   Sean: I prefer to have full word when we can
   ... it would be cleaner to have a full word

   David: don't have a strong feeling. it's easier if it's full word
   ... in this case, it's in metadata namespace.

   Glenn: SMIL, SMIL and XHTML are all consistent in using desc

   Resolution: ISSUE-7 is closed with no change


   skipped (needs Geoff around)


   skipped (needs Geoff around)

   Glenn: who proposed it?

   Sean: I did

   Glenn: we made design choice not to allow linking to external media
   ... we had a long discussion about that a while ago, fonts, images,
   audio, ... we decided we did not want that

   Sean: I'm doing it in a private namespace right now, so not an issue

   Resolution: ISSUE-10 is postponed to v.next


   Resolution: we'll make the change

   ACTION: Glenn to fix inconsistency with regard to time container
   defaults [recorded in


   Glenn: I could add an editorial note to clarify the semantic from

   ACTION: Glenn add an editorial note to clarify the semantic from
   SMIL for dur and end [recorded in


   Resolution: ISSUE-19 closed. no change.


   David: that seems late to make a change now

   Sean: we may reopne them if it's a major problem
   ... happy to postpone to v.next. ditto for ISSUE-21, ISSUE-22,
   ISSUE-23, ISSUE-24, ISSUE-25

   ACTION: Philippe to foward CEA communication to the list, if ok
   [recorded in

   Sean: check to make sure it's ok to make it public


   Glenn: [...] it's specified in the definition of overflow in the
   ... [citing the spec]

   Sean: we will remove that if we don't have dynamic flow?

   Glenn: yes

   Resolution: ISSUE-26 is closed. no change.


   Glenn: we have some inconsistency indeed with XSL...
   ... looks like we tried to make it more consistent with naming
   instead of XSL

   Sean: everybody said we should be consistent with XSL or we should
   document why not

   Glenn: I'm happy to make it consistent with XSL

   Resolution: ISSUE-27 to make it consistent with XSL (already done)

Next steps

   Philippe: I'll look into creating a new questionnaire with new

Namespace style

   Philippe: we have a namespace inconsistency between the Adobe
   implementation and the latest version of spec (November 2006).

   Glenn: we never fixed the namespace in the spec

   Philippe: other groups usually fix it at CR by respect for
   implementators, unless they make significant changes during CR.

   Sean: we can be lenient for now in the test suite and fix it later.
   Our immediate goal is to test the features.

   Philippe: ok. as long as we demonstrate interop at the end of CR,
   i'm ok to be lenient for now. I'll use the namespace supported by
   Adobe and WGBH in the tests for now. For end of CR, we have 3
   choices: keep the Adobe one, use the November 2006 one, or get a new
   one. For the new one, we could use a short version for namespace
   btw, e.g. [24]http://www.w3.org/ns/dfxp. It's easier to remember
   since you don't have year/month.

     [24] http://www.w3.org/ns/dfxp

   Sean: would very much like to switch to that indeed

   Resolution: we'll use a short namespace (ie no year/month) by the
   end of CR. Namespaces will remain a moving target for now.


Summary of Action Items

   [NEW] ACTION: Glenn add an editorial note to clarify the semantic
   from SMIL for dur and end [recorded in
   [NEW] ACTION: Glenn to fix inconsistency with regard to time
   container defaults [recorded in
   [NEW] ACTION: Philippe to foward CEA communication to the list
   [recorded in

   [End of minutes]

Received on Monday, 26 January 2009 22:49:36 UTC