WAI CG special discussion on HTML alt

18 Feb 2009


See also: IRC log


Cooper, Gregory_Rosmaita, Janina, Joshue, Gez_Lemon, Matt_May, Steve_Faulkner, Gregg_Vanderheiden, Ben_Caldwell, Jan_Richards, Janina_Sajka




<oeddie> janina, you need to have zakim dial you again - wait MC is calling you

<oeddie> chair: Janina_Sajka

Today's Agenda

<oeddie> JS: some discussion in PF HTML5 Issue caucus call on friday about @longdesc

<oeddie> http://esw.w3.org/topic/PF/XTech/HTML5

<oeddie> http://esw.w3.org/topic/PF/XTech/HTML5/Caucus

<scribe> scribe: MichaelC

<oeddie> JS: automated image analysis hueristics canard

<oeddie> http://esw.w3.org/topic/PF/XTech/HTML5/MediaSpecificElements

<oeddie> (contains proposed mapping for @alt to LEGEND and @longdesc to proposed DESC child of media specific elements

<Joshue> JOC: It would be a great idea to have something like that to go to the HTML 5 WG with.


<oeddie> http://www.w3.org/2009/02/12-cg-minutes.html

<oeddie> http://www.w3.org/2009/02/04-cg-minutes.html

<oeddie> http://www.w3.org/2009/01/21-cg-minutes

<oeddie> http://esw.w3.org/topic/HTML/Action54AltAttributeThirdDraft

<oeddie> http://www.w3.org/2009/02/13-pf-minutes.html

<oeddie> gregg, that's the last HTML Caucus call

Summary of proposition: http://lists.w3.org/Archives/Member/w3c-wai-cg/2009JanMar/0106.html

<Joshue> +1

RESOLUTION: We want HTML5 to provide mechanisms for providing text alternatives. (discussed 4 February)

<oeddie> plus 1

<oeddie> http://esw.w3.org/topic/HTML/Action54AltAttributeThirdDraft

<oeddie> latest HTML5 draft: http://www.w3.org/TR/html5/

<oeddie> http://www.w3.org/html/wg/markup-spec/

<oeddie> http://www.w3.org/html/wg/tracker/issues/66

<oeddie> http://dev.w3.org/html5/spec/Overview.html#the-img-element

<Zakim> Joshue, you wanted to say that HTML should natively support accessibility

<oeddie> also, native support is part of WAI-PF's first official communique with the HTML WG

<Joshue> zakim. mute me

<oeddie> http://esw.w3.org/topic/HTML/AccessibilityDependencies

<oeddie> http://esw.w3.org/topic/HTML/ARIAIntegration

gv: is ARIA going to be taken into HTML 5?

mc: understand it's an editorial task to just wholesale take it in

<oeddie> http://www.w3.org/html/wg/markup-spec/#aria

<oeddie> http://www.w3.org/html/wg/markup-spec/#common-aria

<oeddie> http://www.w3.org/html/wg/markup-spec/#implicit-aria

<oeddie> http://www.w3.org/html/wg/markup-spec/#aria-attrs-all

gv: concern about source of expertise for accessibility guidance in the HTML spec

<oeddie> that is correct, matt

mm: note HTML WG has dependency on PFWG for accessibility advice; we're just asking them to follow up on this

<oeddie> http://esw.w3.org/topic/HTML/ARIAIntegration

<oeddie> http://esw.w3.org/topic/HTML/AccessibilityDependencies

sf: the examples in HTML draft haven't been produced by WG, just the editor

<oeddie> http://www.w3.org/TR/html5/embedded-content-0.html#the-img-element

<oeddie> http://www.w3.org/TR/html5/embedded-content-0.html#alt

jr: our approach should take into account an expectation that HTML 5 process won't change

<Joshue> Big +1 to Jan

so recommend terse phrases for spec

gv: that should work, as long as nobody decides to require something that goes beyond what we've advised, and have to write it up

<oeddie> GJR notes that the aria integration is part of the prototyped HTML5: The Markup Language and NOT part of 2009-02-12 HTML5 draft (http://www.w3.org/TR/html5/)

<oeddie> move on


<oeddie> http://esw.w3.org/topic/PF/XTech/HTML5/MediaSpecificElements

gr: propose we consider HTML4 @alt like HTML 5 figure/legend, and HTML 4 @longdesc like HTML 5 desc

gv: sounds like an accessibility support issue; those would be sufficient techinques

<oeddie> "Towards A Unified Approach to HTML5's Media Specific Elements, version 0.1"

aria-labelledby and aria-describedby would be alternate techniques, if accessibility supported

straw proposal: longdesc not be included in HTML 5

(supplanted by ARIA)

WCAG would accept it as long as it remains accessibility supported

is it ok for aria-describedby to reference something that itself provides a link to an external description?

<oeddie> 3 points: 1) need for terse and long descriptors; 2) need mapping for backwards compatibility; 3) still unsure if can reference a label or description that is not contained in same document as object being labelled or described by use of the "rel" attribute with the A (anchor) element

gr: <see above note>

concern that aria-describedby referencing a link sounds like a roundabout d-link

<oeddie> we need to mock up tests -- i volunteer

<Joshue> +q

gv: gets around the fact that ARIA doesn't provide a way to reference external description

gr: XHTML 2 WG would like ARIA to support CURIE in ARIA 2

<oeddie> gregg, CURIES: http://www.w3.org/MarkUp/Drafts/Overview.html#curie

<oeddie> http://www.w3.org/MarkUp/2008/ED-curie-20081023

<oeddie> http://esw.w3.org/topic/HTML/LongdescRetention

joc: <comment to follow>

js: sounds like we do want a mechanism to reference a description in an external resource

<Joshue> JOC:If Longdesc content could be buffered by a UA like a screen reader and then activated when the user chooses using a predefined key in the AT. This would be a more elegant solution.

gv: does look like something we overlooked; would hope ARIA would provide this

js/mc: sounds like a LC comment on ARIA

<oeddie> "HTML 5 Issue: Image Equivalent Content" (http://esw.w3.org/topic/HTML/LongdescRetention) - "Requirement: In situations where images are not available to the user (because of disability, choice, or UA limitation) there is a need for a mechanism that presents equivalent content to the user, either as an alternative to the image or in a side-by-side exposition. Equivalent content is not, nor should it be, and either/or proposition, and its method of exposition should

<Joshue> JOC: Could ARIA properties reference external URI's etc in the future?

gv: why doesn't ARIA have external references?

mc: believe it was to make documents self-contained

js: it's clear we want @longdesc-like functionality and might want it in ARIA, notwithstanding that design approach

gr: <see above comment>

<Zakim> oeddie, you wanted to say a lot of thought excercises on long descriptors and exposition methods at http://esw.w3.org/topic/HTML/LongdescRetention

gr: user agent preferences

gv: how does UA know?

absent longdesc?

gr: this is why @longdesc is still part of spec

gv: isn't it obsoleted?

gr: above is one of several competing proposals

attempt is to provide a consistent content model + backward compatibility

uses desc element, which can contain anything

gv: isn't that same as using @aria-describedby?

gr: wanted to cover use case for side-by-side, not just either-or- exposition of alternative content

gv: so for @aria-describedby, shouldn't necessarily jump to it, but be able to pull up its content without moving

if the target is a link, you can just follow it, and UA present it in a side presentation

<oeddie> phone problems

<oeddie> can hear but you can't hear me

gr: <some thing about @src global attribute>

gv: in this proposal, @aria-describedby and desc would behave in similar manner, and would be parallel techinques

<oeddie> describedby and DESC should be indentical content

following behaviour suggested above

<oeddie> agree with GV

then we wouldn't need to add external reference to ARIA

<oeddie> http://esw.w3.org/topic/PF/XTech/HTML5/RoleAttribute

<oeddie> Role module is extensible

sf: target of @aria-describedby could have a role that would guide UA behaviour

jr: agree we don't want new features that are redundant with what we already have

the proposal for how to display things is up to UA

<Zakim> oeddie, you wanted to say there is a proposal in front of XHTML2 and HTML WG to add "for" to the global attribute list for both MLs

let's focus on the information architecture to support

gr: there is a proposal to have @for=IDREF as a global attribute

so UA could handle

gv: ability to invoke description without leaving your place an important aspect of these proposals

<oeddie> use cases: http://esw.w3.org/topic/HTML/LongdescRetention#head-30d2dc4fc6689c2e76c9626f3ea86b9023d8499c

<oeddie> proposed and existing solutions: http://esw.w3.org/topic/HTML/LongdescRetention#head-cebbc654950a1139904c489c40c021ad3a200ddd

<Stevef> in-place long description from another URL could be done using iframe/javascript

looks like we have workable approaches

<janina> +1

therefore circle back to deprecating longdesc, in favour of ARIA

<oeddie> that's a LOT of ifs...

<oeddie> MC: think getting to conclusion; wrap up and comb hair out of cat later

js: don't hear change to the previous resolution about deprecating longdesc

we've clarified need to reference xsd:anyURI somehow

but that seems covered

<greggvanderheiden> If ARIA describedby is adopted by HTML5 and describedby can support a link to content off page then Longdesc should be deprecated or obsoleted.

<Joshue> JOC: thats fine with me.

<Stevef> +1

<JR> +1

<Gez> +1

<oeddie> virtual face2face

<Joshue> +q

<Joshue> -q

RESOLUTION: If ARIA describedby is adopted by HTML5 and describedby can support a link to content off page then Longdesc should be deprecated or obsoleted.

<oeddie> actively discussing @alt - the action to report in HTML WG is on me

<oeddie> i agree with steve

<oeddie> beware of false deadlines and sending mixed messages

<discussion of communicating status to HTML WG>

<Joshue> JOC: I think that the issue is largely a user agent issue.

<Joshue> The language should facilitate semantic and other associations and the rest is up the technology that interacts with it.

<oeddie> good point, JOC

<oeddie> jan, you don't buy the use cases at: http://esw.w3.org/topic/HTML/LongdescRetention#head-30d2dc4fc6689c2e76c9626f3ea86b9023d8499c

<oeddie> viewport

<oeddie> there is also the case where the long descriptor is actually an ARIA-enabled widget -- such as a geneological tree or flow chart

<oeddie> provide same ah-hah! eureka moment through use of a widget as long descriptor for a static flow chart

<oeddie> that's what global for/id associations could do, jan

<oeddie> http://esw.w3.org/topic/PF/XTech/HTML5/Caucus

<oeddie> friday, 20 February 2009 at 10:00 A.M. EST

<oeddie> JS: summary and headers next wednesday; alt on friday

<oeddie> we can preclude that by getting "role" supported in HTML5 - "role" for machine processing, "alt" for human processing

<oeddie> http://esw.w3.org/topic/PF/XTech/HTML5/RoleAttribute

<oeddie> cascade of alternatives through programmatic associations

<oeddie> alt must be neaningful, role value is machine-parsable

<oeddie> alt for humans, role for machine

<oeddie> http://esw.w3.org/topic/PF/XTech/HTML5/RoleAttribute

<oeddie> ACTION: GJR and JR - work on wording for discussion on friday [recorded in http://www.w3.org/2009/02/18-cg-minutes.html#action01]

<trackbot> Sorry, couldn't find user - GJR

Summary of Action Items

[NEW] ACTION: GJR and JR - work on wording for discussion on friday [recorded in http://www.w3.org/2009/02/18-cg-minutes.html#action01]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.133 (CVS log)
$Date: 2009/02/18 21:36:00 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.133  of Date: 2008/01/18 18:48:51  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/elegeant/elegant/
Found Scribe: MichaelC
Inferring ScribeNick: MichaelC
Default Present: Cooper, Gregory_Rosmaita, Janina, Joshue, Gez_Lemon, Matt_May, Steve_Faulkner, Gregg_Vanderheiden, Ben_Caldwell, Jan_Richards, Janina_Sajka
Present: Cooper Gregory_Rosmaita Janina Joshue Gez_Lemon Matt_May Steve_Faulkner Gregg_Vanderheiden Ben_Caldwell Jan_Richards Janina_Sajka
Agenda: http://www.w3.org/mid/20090217184130.GB3809@sonata.rednote.net
Got date from IRC log name: 18 Feb 2009
Guessing minutes URL: http://www.w3.org/2009/02/18-cg-minutes.html
People with action items: gjr

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

[End of scribe.perl diagnostic output]