See also: IRC log
<trackbot> Date: 25 May 2011
<janina> Meeting: HTML-A11Y telecon
<janina> agenda: this
<scribe> scribe: silvia
next item
discussion about what the tracks are
one type is a multi-channel mix in one track, where one channel has speech-only
such a track is an alternative to main audio
another type is an additional track that contains only speech and can be increased in volume over other tracks
proposal is that we may need both
however there is not enough real-world use case knowledge in the group to validate the use case
we will seek further input from experts before making a decision
janina: last week we seemed to have a solution for how to appropriately mark up alternative text, both short and long for video
including poster
both for graphical and non-graphical browsers
I took it to PF today
for example to find out if we retain language markup
and the answer is: yes - the markup is retained for elements, but not for attributes
however, if you go into a non-graphical browsing mode, I'm not sure we are safe
eric: a non-graphical browser would not support the video element and thus should take the fallback inside it
silvia: a lynx developer seemed to indicate that once a non-graphical browser supports html5, it would not display the content inside the video element
instead it would compose such text from the video element's attributes
I am not even sure we should be discussing this here in the a11y TF, but it's an issue for the larger group
eric: it would then also have to support the aria attributes so it should have enough information from aria-describedby as we have suggested it
janina: what are the next steps? need to discuss in the wider community?
I wanted to introduce it into WAI today
sean, eric: seems a fine solution
silvia: the issue that John indicated was that when aria-describedby points to structured markup, screenreaders currently just read it out as text
when it should in fact be read out as structured markup, e.g. including URLs
Janina: yes, PF also agrees with this and screenreaders need to start interpreting such text correctly
next step should be to get this information out more widely so we get good information on that
janina: should this go into the HTML5 spec and / or into the HTML a11y User Agent implementation guide
silvia: there is no spec change necessary for HTML5, but it might be good to add an example into it to get the approach in front of more eyes
eric: what we need to explain to people is how aria-describeby works
<janina> http://www.w3.org/html/wg/wiki/ChangeProposals/ImgElement20100504
<janina> sorry wrong ...
<janina> http://www.w3.org/TR/html-alt-techniques/
it should probably go into the HTML alt techniques document
janina: we also want it in the user agent guide, too
in particular that that level of markup should be supported
e.g. that uris need to be interpreted
eric, silvia: sounds fine by me
janina: will forward it in both cases
silvia: the only thing that we may need a change proposal for, if we still want it, is the @transcript attribute
it could be replaced by a @longdesc attribute as a means to link to a full transcript
sean: I think they are semantically different
silvia: instead we can just have it inside the aria-describedby markup just like we have it for the poster long description
<video src="file.mp4" poster="poster.png" aria-describedby="posteralt videosummary" aria-label="MyPlayer video player">
<p id="videosummary">A Clockwork Orange Trailer (<a href="transcript.html">Transcript</a>)</p>
<p id="posteralt">Poster frame is a clockwork orange movie poster (<a href="posterlongdesc.html">long description</a>)</p>
<p> <a href="file.mp4">Download the video file</a></p>
</video>
there is a lot of indirection now
sean: is the fallback content even in the dom?
eric, silvia: yes it is
sean: with this markup, there is no semantic markup that a machine can rely on to interprete as the transcript
the user has no way to get at that content
silvia: yes, an attribute @transcript would allow creating a context menu item for users
so do we want to start a discussion in the larger HTML5 group about introducing such an attribute?
I can start such an email discussion on public-html
janina: yes
… general agreement
<scribe> ACTION: janina to work with authors of HTML alt techniques document and HTML a11y user agent implementation guide to add the aria-describedby video example there [recorded in http://www.w3.org/2011/05/25-html-a11y-minutes.html#action01]
<trackbot> Created ACTION-127 - Work with authors of HTML alt techniques document and HTML a11y user agent implementation guide to add the aria-describedby video example there [on Janina Sajka - due 2011-06-01].
<scribe> ACTION: silvia to follow-up on @transcript attribute [recorded in http://www.w3.org/2011/05/25-html-a11y-minutes.html#action02]
<trackbot> Created ACTION-128 - Follow-up on @transcript attribute [on Silvia Pfeiffer - due 2011-06-01].
Judy will follow up on a reply to the ISO/MPEG group
re kind attribute values
other actions are still open and continue
there is also a bug on providing track usage examples - silvia will follow up
note that Google proposed a WebVTT WG charter to W3C
janina: we still have to take up the topic of navigation
silvia: had a brief discussion with Ian about this and we may be able to satisfy it with existing techniques - have to look at it in more depth - there's a bug in the HTML bug tracker
janina: keen to look at the Web Audio API
there is an ETSI standard on this for broadcasting
This is scribe.perl Revision: 1.136 of Date: 2011/05/12 12:01:43 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Found Scribe: silvia Inferring ScribeNick: silvia Default Present: Janina, Sean, silvia, Judy, Eric, Plh Present: Janina Sean silvia Judy Eric Plh Regrets: John_Foliot Marc_Watson Found Date: 25 May 2011 Guessing minutes URL: http://www.w3.org/2011/05/25-html-a11y-minutes.html People with action items: janina silvia WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]