W3C home > Mailing lists > Public > public-html-a11y@w3.org > April 2011

[minutes] Media Sub-Team meeting, April 20

From: John Foliot <jfoliot@stanford.edu>
Date: Wed, 20 Apr 2011 17:20:39 -0700 (PDT)
To: "'HTML Accessibility Task Force'" <public-html-a11y@w3.org>
Message-ID: <032601cbffb9$f0f9da90$d2ed8fb0$@edu>

Minutes from the April 20 2011 HTML Accessibility Task Force Media
sub-Team Weekly teleconference are now available:


They are also available as plain text following this announcement -- as
usual, please report any errors, omissions, mis-attributions,
clarifications and the like by replying-to this announcement on-list.


Media Subteam 20 April 2011

20 Apr 2011

   See also: [2]IRC log

      [2] http://www.w3.org/2011/04/20-media-irc


          Bob, Eric_Carlson, Frank, John_Foliot, Judy, Sean, mark,




     * [3]Topics
         1. [4]Identify Scribe
         2. [5]Issue-152 Multitrack
     * [6]Summary of Action Items

Identify Scribe

   <judy> Sean will scribe, Judy will fill in when Sean is commenting

   <scribe> scribe: Sean

Issue-152 Multitrack


   Silvia and JF catching up on email

   Ian has introduced some changes

   yes please

   Judy recommends that we maintain formal and informal communication

   Judy: Good to record out discussion

   Frank: On track to have a single coherent proposal, would like our
   CP to be against whats in the spec, rather than a CP against a CP

   to consolidate into one single change

   JF: if there is still difference between this group and Ian, how do
   we capture that

   we should not let anything slip

   Mark: arent we required to produce a CP against the current W3C spec

   <silvia> +q

   Judy, yes; typically what we do against a rolling editors vresion is
   comment wrt to a time stamped version

   Silvia. the Wiki page is a formal record to some extent

   have removed things as Ian has added them

   Ian's CP is in the WHATWG spec

   ours is in the Wiki

   the changes are not uncontrolled

   SJudy: That can be tricky

   multiple moving targets dont give the record trail

   unless you dive into the histories of the individual pages

   email also gives some trail

   we need to send things that are static



   Silivia: working to that

   see urL

   ask to work through the 'silvia's notes section

   this is the proposal we are going to submit

   work through the 7 points

   and Silvia can prepare the email

   for dicussion

   all: general agreement

   point 1.

   make track more like text track, now recorded as a bug

   Silvia - tracklist has a set of getXX functions

   these are different to how text track is defined

   where they are attributes

   Philip suggest making a list of object with attributes for

   we dont need to record this as its a separate bug

   JF: Q: What are the pros and cons

   if it goes after LC

   Eric. the issue is that philip has already filed a bug, and its just
   a change to the interface

   we dont need to restate it

   <silvia> [9]http://www.w3.org/Bugs/Public/show_bug.cgi?id=12530

      [9] http://www.w3.org/Bugs/Public/show_bug.cgi?id=12530

   move on and leave point 1

   Item 2.

   it should include id, label and kind

   JF: Ian was unclear about kind

   what are we discussing?

   Silvia its about adding ID and kind

   to get a unique identifier to create a fragment url

   Sean: should also have a getFragmentIRL function

   Siliva: agree, but ID is still needed

   Sean: have both

   Eric: prefer to not have the function

   problem is that there are files for which the URL for the track is
   not the same as the parent movie

   which is another file on the server

   asking a track for a URL, if its a reference to a different file;
   this needs to be resolved

   it is also a problem that the fragment spec is not final

   it seems to be something that could be done in a JS library/

   Silvia can find outr which tracks exist

   composing the fragment url to combine in one URL

   Sean: do we need additional information?

   Eric: Asking a track for a reference to itself

   refering to the track from the outside is unambiguous

   from the inside there is a possibility of confusion

   defintion of the attribute needs to say what it refers to

   Silvia getID we definitley want

   get URL may be a convenience

   we dont need to agree on it now to resolve 152

   we need to look at the current src to see if its adequate prefix to
   a #id

   <JF> +1 to @kind

   Silvia: kind is required to discover what the content is intended

   <frankolivier> +1 to @kind

   all: agree kind is necessary

   Eric: one problem is how we define the media information to a kind

   doesnt exist in in MP4 and Quicktime

   we are working on a proposal

   but hasnt been defined yet

   secondly, no authoring tools to let you add it

   maybe some command line tools

   so it will be a while before this data shows up in files

   but agree its important, but will be difficult for a while

   Silvia: need it in the spec for interoperability

   JF: if the additional data is in the wrapper, how is it exposed to
   the player

   SIlvia: is the data- attribute appropriate?

   eric: we need a mechanism to express a kind of an inband track

   this is not on the markup

   its in the media

   Eric we have a bit of chicken and egg problem, but it will get

   Silivia: there are slots in media files, but we need time to
   establish usage

   Eric: there are fields in Quicktime files, but not standard values

   mark: in mpeg-DASH they have been looking at this, and will add the
   necessary annotations

   JF: so we do need @kind

   did we not have other kinds

   Silvia: thise were for text

   this is AV

   we need to think about other kinds

   sean: clear audio sould be a kind

   speech only

   JF: lang, ISO language codes, there doesnt seem to be consistency

   do we need to specify language codes for sign

   Silvia: yes the latest ones are fairly complete for sign

   so not a problem

   Mark: are there others like no repetetive stimulis


   Eric: that's not really a kind, but a characteristic of the asset

   so slightly orthogonal issues

   JF: +1

   these issues need to be considered, but its not an @kind value

   Eric: D Singer and I have never gotten any traction on it

   we cant do it in the next 48 hrs

   JF: do we have a final list

   Sean: should this be an open ended list

   Silvia: yes

   JF: agree

   Judy: can you restate the issue

   eric: there is a need for content authors to decalre tha
   accessiility reqs that content meets, but also markup content to let
   the UA pick the content the user does not want

   do the UAAG WG have anything relevant

   Eric: the issue is that the content author needs to be able to
   describe the content

   Mark: MPEG-DASH is defining these roles, and sent a liaason on this
   question and multiple roles can be assigned

   Judy: can I be put on that htread

   Mark: the liason should come formally soon

   it was addressed to this group

   Silvia: I think there are some open issues, but we can maybe pursue
   them separately

   maybe as a media query

   Judy: the restatement was helpful

   can you send the materials

   Silvia: added clear audio

   JF: done with kind

   next item, loop

   Silvia: loop is currently disabled for a combined set of elements

   we have discussed how it should work

   maybe we can leave it for now and discuss with ian

   do we actually want loop to be solved

   JF: straw poll of opinion

   Eric: we should hhave it

   Frank: agree

   Sean: no reason not to

   mark, Judy: no opinio

   consensus we should have it

   net item : autoplay

   Judy can we be clear on the question

   eric: should we ever allow autoplay for a group of media resources

   silvia: we have it in the requirements

   its done by single and inband multi track

   but not by slaved elements

   do we want it

   Judy: current wording?

   JF: is selective switching possible?

   o.html#media-controllers - you need to scroll down to where
   "autoplay" is mentioned


   Judy: its a distraction issue, and its not always easy to figure out
   how to switch it off

   Eric: a non issue as it would be foolish to have separate ungrouped
   elements with media that should be synchronised

   you always want them to play in sync

   its just a question of enablement

   JF: need to move on

   do we have a solution today?

   Eric: no

   its about user agent preferences

   <silvia> A
   o.html#mediacontroller is a blocked media controller if the
   o.html#mediacontroller is a
   o.html#paused-media-controller, or if any of its

     [15] http://www.whatwg.org/specs/web-apps/c

   Judy: need to defer this and go to the other issues

   item: some attributes are missing in the controller

   Silvia. we had request to add some events, but these relate to the
   readyState, so we need to decide whether we need a combined

   it seems readystate is inconsistent

   so perhaps we dont really need it,

   but we can wait till the result of the bug

   Eric: i think its useful

   on the media controller, its not possible to find out the combined
   state otherwise

   Frank: agree with Eric, but need more discussion

   JF: does it go in the CP?

   Frank: could it go in as a bug outside the CP

   JF: not sure,

   easier to delete it later

   general consensus: should go on the controller

   Silvia: will add to the CP

   item 7


   ended state

   Silvia: : the onended event is not included, as Ian not clear what
   it means

   think it is necessary

   Eric: ended is required

   Bob: agree, Mark: no opinion

   consensus : keep it in

   item 7

   how to create controls for the combined grou

   Silvia: how are the automatic controls created for a combined

   Ian seems to thin its possible on each video element, and they will
   control the controller

   this seems like an OK solution

   its probably not possible to create one over multiple elements

   Eric: my interpretation is that if show for any element, it works
   for all the elements

   from a display point of view they will cover only one video element

   JF: so this is a solved problem

   return to autoplay

   Judy: can we restate the question

   silvia: should the controller have an autoplay idl attribute
   representing all the slaves

   Judy: I think so

   <judy> Sean: concerned that we are not clear about the semantics of
   the idl attribute

   <judy> ...will it be controlling all of the elements, or only the

   <judy> [others had stated that they support idl, or that they likely

   <frankolivier> (Need to drop for another meeting)

   <judy> [discussion about (un)clarity in semantics of autoplay, and
   how that relates to the question on which we are trying to agree on
   a position ]

   <judy> bob: the behavior on autoplay should be the same as it would
   be if the media player were invoked manually

   <judy> several: need communication with ian

   <judy> silvia: writing him a letter

   <judy> john: we seem to agree on what's needed

   <judy> ...and we've gone through all 7 of silvia's bullet points

   <judy> silvia: yup, have my notes, will draft letter, circulate it
   within the next hour

   <judy> john: let's agree to state that as a consensus of the
   accessibility sub-group

   <judy> ...let's give people 24 hours to agree on that

   <judy> ...any tweaks needed, ask silvia to help with

   <judy> ...for all, please send a +1 back to the list

   <judy> ...thank you everyone so much, good process, good effort

   <judy> judy: thanks all

   <JF> hi friend

   <JF> so the only thing i have not done is tell rrsagent to p art

   <JF> scared to do so

Summary of Action Items

   [End of minutes]

John  Foliot
Program Manager
Stanford Online Accessibility Program
Stanford University
Tel: 650-468-5785

Soap Is a program directed by the
Vice Provost for Student Affairs
Received on Thursday, 21 April 2011 00:21:08 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:05:19 UTC