W3C

- DRAFT -

HTML Accessibility Task Force Teleconference

17 Nov 2011

See also: IRC log

Attendees

Present
Mike_Smith, Janina_Sajka, John_Foliot, Léonie_Watson, Rich_Schwerdtfeger, Cynthia_Shelley, Paul_Cotton
Regrets
Chair
Janina_Sajka
Scribe
Léonie Watson, MikeSmith

Contents


<trackbot> Date: 17 November 2011

<MichaelC> Meeting: HTML-A11Y Task Force Teleconference

<MikeSmith> chair: Mike

<scribe> Scribe: Léonie Watson

<MichaelC> scribeNick: lwatson

Subteam Reports: Bug Triage; Canvas; ARIA Mapping; Text; Media

RS: The only thing issue 131 needs to be opened up for is text baseline.

JF: With Web VTT, it's been moved into two community groups. One looking at Web VTT and one looking at Time Text.
... One possibility is that the two groups could merge.
... The Web VTT group is looking at mapping between Web VTT and Text Track.
... I think Web VTT is likely to become the default format.
... TTML will map over to Web VTT and vice versa, and browsers will natively support Web VTT.
... No guarantee, just my assessment.

RS: Will they do something formal with the community group?

JF: The Web VTT group is chaired by David Singer. When we can say of Web VTT that it solves all the problems we have, then it's likely the community group will come forward to be formalised in one way or another.

RS: We're trying to decide whether to ask one of the IBM team to join the community group. What's the meachnism? Is it worth us doing that?

JF: Yes, I believe so. David or Silvia are the people to approach.
... There are many implementations of Web VTT in the wild. I can send you some information off list.
... I think it's the DFX profile that IE supports.
... The spec says that this is handled by track. We're handling a file format to the browser with a file format that isn't mentioned anywhere in the HTML5 spec.

<paulc> http://blogs.msdn.com/b/ie/archive/2011/10/12/html5-video-captioning.aspx

MS: TTML doesn't use CSS for styling...

JF: It's not as native an issue as people think. You're right, an an XML language TTML can't be styled.

<paulc> The above is a prototype.

JF: Web VTT can be styled.
... Right now the HTML5 spec just says "a timestamp format" and steps out of the way.

PC: I posted a link to the MS blog about the prototype.

MS: Any more on media John?

JF: The sub team needs to meet again.
... There are some @kind issues which I think we need to track.

PC: We had some good dialogue about canvas at the F2F. How is this being progressed?

RS: We need to have hit testing in place. We have agreement from people including Frank, Jonas and others. Frank is going to write the hit testing spec.

<paulc> W3C first developers conference: http://w3conf.org/

RS: I'm trying to find out if we can leave the draw focus ring as it is, and look at text baseline on its own.

LW: We've carried out a review of all previously tagged bugs, to confirm everything was...

JS: No meeting of the text sub team this week.

MS: ARIA mappings?

CS: We didn't hhave a meeting this week. It will be rescheduled after Thanksgiving.

JS: Paul, we talked last week about letting the chairs know the TF will respond on longdesc. I wanted to confirm for clarity. Are we looking for a window here, or a survey? Is everything clear?

I think TPAC was helpful in moving this along. I want to make sure we keep building on that. Are we good to go all around on this? Also, where in the process are we?

<richardschwerdtfe> dropping off

MS: Paul has dropped. We'll follow this offline.

Other Business

JF: Would an update on poster alt be helpful?

RS: This would need to be in ARIA 1.1. We need to get 1.0 into LC before we can do anything. I think a way to move it along would be to get browsers to implement it.
... We'd need a video role for the poster, then some kind of attribute (do we have a name?) that would take a URL for a verbose description.

JF: Correct.
... The idea is to call it aria-poster.
... In some instances holding images are being used for things other than video. For example an online survey created with HTML5 form elements, but behind the form is a hand drawn survey of questions.
... There is a dis-association between the image and the content. We can probably exapnd the use case for aria-poster to include these instances perhaps.
... If we think of it as a background image that the video is on top of (rather than a foreground image that the video is underneath), we can broaden its use.
... Early conversations with Frank and others suggested general agreement.

RS: So what would you call it then?
... Something like aria-background?

JF: Yes, something like that.
... There would need to be some semantic linking, but if we think of it as a background image it could be applied to any kind of semantic container.

RS: Would you want it to take a URI?

JF: I think that's the only way to support rich text and internationalisation.

CS: If you have aria-background-image pointing to a URI there's a likelyhood that developers will think it's the URI for the image.

JF: I don't mind what we call it.

CS: Having a text alternative for background images works for screen readers, but they're problematic for people who need high contrast. We don't want to encourage people to use background images for important content.

JF: The point I tried to bring forward about discoverability and usability is a reuirement. It's a requirement for user agents.

CS: If you expand the use case to background images, it's encouraging poor practice.

JF: Don't disagree, but it's happening out in the wild at the moment already.

CS@ Whether we soleve this through background images in terms of colour contrast, or we solve it here, we need to solve it.

RS: It sounds like you want some kind of icon description John?

JF: No, that's too confusing.
... We have an image painted to the screen before a video is presented.
... We need a way to describe that image.

RS: So when the element has a adescription and the associated image has relevant information.
... So a person uses a background image so it lines up with foreground content. There is a disconnect there. There's a reason WCAG 2.0 has these requirements.

<MikeSmith> scribe: MikeSmith

[John describes use case]

Cynthia: sounds like that graphic is decorative
... and that there are text labels on it that constitute the actual UI

JF: I re-think the idea of poster image as being a background image that's behind the video, just before it plays

<lwatson> scribe: Léonie Watson

<scribe> scribenick: lwatson

JS: It's going to be a highly used situation

JF: Question is what I should be doing next?

RS: Give us a name and work with us to get support in the browsers.
... Typically accessibility APIs don't have multiple long descriptions. Question to the browser vendors is, would they be willing to render this into a window that can be viewed?

JF: Ok, I'll work on this.

MS: Any other business?
... Scribe for 1st Dec?

JF: Put me down.

<MikeSmith> no call next week

JS: No call next week. US Thanksgiving holiday.

<scribe> Chair: Mike Smith

<scribe> ...done correctly. Next we'll be pulling together lists assigned to specific sub teams where more information is needed. We'll send these to sub team chairs and invite them to attend a triage call if they think there are ways the triage team can support them./

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.136 (CVS log)
$Date: 2011/11/17 16:55:55 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
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)

Succeeded: s/PC: TTML/MS: TTML/
Succeeded: s/be come/be some/
Succeeded: s/LW: We've done a check of all tagged bugs. We're now pulling out lists of bugs needing more info by specific sub teams. We'll include an invitation for the triage team to help go through the bugs and extend an invitation for sub team chairs to join a triage call if it would be helpful to asign work./LW: We've carried out a review of all previously tagged bugs, to confirm everything was.../
Found Scribe: Léonie Watson
Found ScribeNick: lwatson
Found Scribe: MikeSmith
Inferring ScribeNick: MikeSmith
Found Scribe: Léonie Watson
Found ScribeNick: lwatson
Scribes: Léonie Watson, MikeSmith
ScribeNicks: lwatson, MikeSmith
Present: Mike_Smith Janina_Sajka John_Foliot Léonie_Watson Rich_Schwerdtfeger Cynthia_Shelley Paul_Cotton
Found Date: 17 Nov 2011
Guessing minutes URL: http://www.w3.org/2011/11/17-html-a11y-minutes.html
People with action items: 

[End of scribe.perl diagnostic output]