HTML A11Y Text Alternatives Sub-Group Teleconference

02 May 2011


See also: IRC log


Cynthia_Shelly, Geoff, Gregory_Rosmaita, John_Foliot, Judy, Laura_Carlson, Leonie_Watson, Michael_Cooper, Rich, Stevef, janina
Lynn_Holdsworth, Marco_Ranon
rich_schwerdtfeger, Rich


<judy> scribe: rich_schwerdtfeger

<richardschwerdtfe> scribe: Rich

<oedipus> scribenick: rechardschwerdtfe

<oedipus> scribenick: richardschwerdtfe


<judy> http://www.w3.org/2011/04/25-text-minutes.html#ActionSummary

Action Item Review

judy: this is the action item collection from last week

<oedipus> http://www.w3.org/WAI/PF/HTML/track/products/5

Clarification mail on summary

<oedipus> Text "product" Tracker: http://www.w3.org/WAI/PF/HTML/track/products/5

judy: janina, would you be available to meet with me to discuss the call

janina: yes, later today

judy: gregory is working on a draft
... judy and shawn are working on a poster
... John, do you have a draft ready

<oedipus> http://www.w3.org/WAI/PF/HTML/track/actions/124

JF: on my things to do this week Judy

judy: we can take the table summary first

<oedipus> @summary draft action item: http://www.w3.org/WAI/PF/HTML/track/actions/125

<oedipus> figcaption and @alt: http://www.w3.org/WAI/PF/HTML/track/actions/121

Action item figcaption and the alt discussion

judy: I updated the section on figcaption

<oedipus> updated figcaption and alt post (judy) http://lists.w3.org/Archives/Public/public-html-a11y/2011Apr/0423.html

judy: I think that action item is done.

<oedipus> role presentation action item (rich and steve) http://www.w3.org/WAI/PF/HTML/track/actions/122

role="presentation" rich and Steve

<inserted> scribenick: oedipus

RS: posted text to list -- read all of maciej's review
... not brought up was native host lang semantics for HTML5 @alt how impacts a11y api mapping -- if not identical, problemmatic
... issue from chairs' review -- was not made clear that reason need to use iinterchangeably (diff from html4) -- alt="" is equivalent of role="presentation"

<Stevef> http://lists.w3.org/Archives/Public/public-html-a11y/2011May/0008.html

RS: host language defines a11y api mapping -- didn't specifically say in HTML4 that image object is removed from a11y API tree

CS: point of contention -- IE doesn't do that and doesn't think it should

<janina> Rich's post is at:

<janina> http://lists.w3.org/Archives/Public/public-html-a11y/2011May/0008.html

Change Proposal Alt text

<inserted> scribenick: richardschwerdtfe

judy: this is an email from Sam
... on Friday Paul Cotton sent an email asking what we are trying to do?

<oedipus> paulc query: http://lists.w3.org/Archives/Public/public-html-a11y/2011Apr/0415.html

judy: We are preparing an email and depending on the response to those we will create a formal objection with expedited appeals
... I prefer we coordinate together on this
... Sam replied, in the shared interest of expediency and you are creating clarification please produce a change proposal at the same time.
... this will save us some time
... Sam then sent another mail saying that he would also recommend that instead of calling our responses clarification emails then we would be stating we are bringing new information
... Janina and I also mentioned our reply to Paul that we were hoping to work with the chairs on this.
... Are there questions so far on this?

<oedipus> SamRuby: "To help speed things up, the response undoubtedly would be that all requests to reopen an issue need to be accompanied by a change proposal: http://www.w3.org/Bugs/Public/show_bug.cgi?id=11447#c3" from http://lists.w3.org/Archives/Public/public-html-a11y/2011Apr/0421.html

judy: he was saying that if you do go the new information route explain that it is new information.
... any questions?

janina: you can't just say they did not have all the facts
... they did not make the correct decision based on the facts they had

cynthia: that is an argument you can't win whether you are right or not
... there was a point where Paul was a bit surprised

<oedipus> RS: function of title and alt text being different? that was what paul identified as "new information"

rich: yes that is the difference between alt and title

judy: it did not occur to us that those making the decision were aware of the fundamental differences between alt and title
... is it ok if we have a mix of clarification and new information?

<oedipus> JF: what may be clarification for us is new info for others

john: ok.

Resolution: no objection to doing a mix of clarification and new information

Reminder of change proposal format http://dev.w3.org/html5/decision-policy/decision-policy.html#change-proposal

scribe; rich

Judy: i think we can deal with each of these items briefly
... there are 4 sub bullets that can be applied without ambiguity
... I am assuming that we would be doing either the first or the third bullet
... we need to make the chairs' lives simple
... any additional thoughts or concerns?
... some of these are not ready
... we have a suggestion from Sam on the list that we add change proposals to clarify what we are saying in our clarification emails.
... any objections to doing that in a one step process?

jf: I do have a question about alt text. I don't know if it makes sense to disambiguate that.

judy: that is agenda item 5
... agenda item 5 has to do with some of our clarification mails to proceed at different paces.
... we have different shared rationals

<oedipus> plus 1 to saving time

judy: for the clarification mails that are ready any objection to doing this as a one step process: (clarification plus change proposals)

Resolution: no objection to having a one step process that include a clariification plus change proposal together

Nearing consensus on updated clarification on alt validation? If not what else needed? http://lists.w3.org/Archives/Public/public-html-a11y/2011Apr/0423.html

Are we reaching concensus on alt validation?

judy: I think that we may have some hefty topics to discuss among those.
... there are notes that have not been rolled into this yet. I suggest we walk through it section by section. There are 4 sections of the 6 subdecisions on this.
... we only have issues with 4 or less of the decisions
... let's walk through these one by one
... The first piece of this is on role="presentation" does not make missing alt conforming.
... there was replaced email on this
... there were questions that leif asked
... if you look at this section of the text, there was an initial block of text that was representative of text between rich and john and there was a comment from Leif

john: I think it was more rich and steve were working on this

judy: we have a short window on this

<inserted> scribenick: oedipus

JB: may have short window -- don't want to let questions linger

RS: talked to SteveF about this -- what hadn't looked at was the a11y api mappings are such that if have alt="" equal semantics for role="presentation" -- in ARIA removes object from API tree -- thought same for alt="", but now Cyns says microsoft has problem with that

CS: general discomfort with removing things from tree

SF: doesn't remove anything from tree if role="presentation"

RS: what do with layout tables? mark with role="presentation"?

SF: not surprised that alt="" is not removed from A11y tree -- in WAI-ARIA does remove stuff from tree when marked role="presentation"

CS: MS never thought removing from tree a good idea
... not 100% positive -- will check

RS: exception in IE news to me

SF: role="presentation" on image removes img from a11y tree
... does alt="" remove from a11y tree -- i would say "no" -- hasn't been implemented yet

JF: why does MS consider this a bad idea?

CS: removing things from tree that complicates things in variety of scenarios

JB: may be something need to deal with anyway
... focus on what we want to say -- clarification or new info?

SF: can confirm that role="presentation" to img element removes from a11y tree

JS: what about legacy viewing HTML5 page?

SF: if UA HTML5-compatible, won't be in a11y tree

JF: what happens if img src="foo.jpg" role="presentation" what happens in IE8?

JS: IE6?

CS: IE8 supports aria -- IE6 doesn't -- get image with src and no alt -- if AT goes through DOM, AT can grab role="presentation" if use DOM-aware AT

SF: for UAs that don't support ARIA would be ignore alt=""

JF: if image used inside link, not going to have role=presentation

SF: right -- no downside then for older UAs if AT uses hueristics to filter out images without alt
... if no alt attribute, will ignore unless user sets to read all images

CS: role="presentation" not there for older UAs

JS: strenghtens our argument
... alt="" redundant and unnecessary

CS: MS says "use both"

SF: main argument against role=presentation is in GUI UAs do something different if alt="" and IMG without alt text

<JF> +q

SF: alt="" is a flag to some UAs to render image differently than when there is no alt text provided by author

JB: like to move towards action item to get this written up

CS: is it really so terrible to require both alt="" and role="presentation"

JF: implementation issue with browsers according to SF -- their problem, not ours

SF: our problems are their problems -- assuming that ARIA should not affect layering violations -- don't want to use ARIA to fix for non-api a11y stuff

<judy> [judy suggests that we present the browser implementation issues/differences as new information, within our clarification]

<Leonie_Watson> JB: Would someone who has been following this discussion be willing to write this up?

<inserted> scribenick: JF

RS: not aware that browser vendors do not want to provide accommodation to aria for renedering

<Leonie_Watson> /me No problem. Over to John...

SF: some browsers will render the lack of image different

if uses alt="" won't show anuy image at all, if they omit then browser shows that an image would exist

this is how the browsers do things - alt="" is a flag to make image visible or not visible

theydon't want to make role=presentation equal the same as ""

SF don't want to base grafical rendering on ARIA Roles

SF: it goes back to the layering issue

they don't want ARIA to use affect anything but a11y APIs

<janina> "layering violation" is a fancy term for "not invented here"

RS: to be honest, we didn't have these concepts 10 yrs ago, so using alt="" was the best we had then

this is no longer a valid justification: ARIA has a section in HTML5

JB: we have a lot to still cover - can we move this to the list?

RS: there is a F2F this week

JB: cyns and rich to continue discussing

<Stevef> text of decision for role=presentation http://www.paciellogroup.com/blog/2011/04/html5-accessibility-chops-the-alt-decision/#d2

<judy> We need to recheck this mail http://lists.w3.org/Archives/Public/public-html/2011Apr/0451.html within this section == Should it be permitted to omit alt when role=presentation is specified? ==

JB: need to revies maciej 's note

are there other issues that need to be discussed

<Stevef> http://www.paciellogroup.com/blog/2011/04/html5-accessibility-chops-the-alt-decision/#d2

SF: point out that reformatted decision as html rich as blog post

<judy> [judy thanks steve and asks if Rich and Cynthia please focus their attention on Steve's text rather than my mail]

JB: asks that rich and cyns look at Steve's text when discussing this iissue
... 1 more question on role="presentation"

if by next monday we get more clarification and consensus, what about the next steps towards Change proposal?

one item is actual spec text

how hard would that be?

SF: good question - for the @title issue it was pretty straightforward

JB: so for @ title it would only take a few hours to write?

SF: yes

JB: for the role="pres" is there anyone else who could take this as a work item?
... could SF take this on as well?

SF: yes


CS: Did we take the time to decide if any of these are worth not fighting for?

JB: yes, we did, and are taking this on 2 levels

we looked at the initial 6

we believe that this was one that was significant to address

it may not require a 'fight' but rather clarrification

figcaption might be more significant

<oedipus> scribenick: oedipus

JB: want to review next section of draft text from Judy

<judy> http://lists.w3.org/Archives/Public/public-html-a11y/2011Apr/0423.html

JB: straw proposal to get a lot of discussion going

<richardschwerdtfe> scribe: rich

JB: second section to consider is meta name="generator"

<judy> look for == On the Co-Chairs' decision on meta name=generator ==

<inserted> scribenick: richardschwerdtfe

judy: on the second section on meta generator
... there has been updated text from Leif who was going to try and join today
... there was a bunch more discussion
... could anyone describe where we are on that

JF: I have got slammed last week. there has been discussion but no meeting of minds
... if I am to understand gregory's point

judy: my core question is how essential is the core issue that leif and benjamin are generating

<JF> +1 to SF

steve: I had not looked at the generator question much

steve; this generator flag is a flag that is used by just under a third of web content and it is not used for that purpose. Unless they get rid of the flag. ...

<oedipus> <meta generator="Gregory J. Rosmaita" ...>

jf: to do this it shuts up the validators

steve: they are using this flag that up to now it shuts up the generators for alt
... nobody bought into that
... we are talking about millions of pages

cynthia: if they want a magic token they should make a new one

judy: it does not weaken our interest in restoring this

steve: it may be unlikely that they are going to take it away altogether

<janina> +1

<Laura> hi

steve: software has to opt in to this. Why would you have a general meta flag rather than stating you just don't want to validate alt?
... this is the only flag that stops validation

judy: on this issue of the meta generator how do we zero in on a solution for this?
... is this something that if you chat with Leif you could zero in on the issue

<JF> +Q

judy: we don't want to boil the ocean

steve: to me these are valid arguments for what is there being tossed out
... John would be a better person to conduct it with

janina: it seems to me that our issue is that we don't provide a correct alt on images
... it is being treated as a get out of jail free card

<Leonie_Watson> +1 to Janina

<oedipus> plus 1 to JS

john: I will spend some time this week. The text Leif has been put together .... it has become verbose
... I will work with Leif to bring this to s discussion page that we can work off of
... you are not in that face to face next week. This has to do with I am not in PF yet
... I will meet janina at the airport tomorrow

cynthia: there are 2 scenarios. email and ...

<oedipus> limitations of tools should NOT shape HTML5

<JF> +Q

cynthia: I don't see the need to validate email

judy: there is a use case coming

rich: the use case is coming

<oedipus> private verus public email -- can ascertain use case for public email (emessages from gov't, orgs, businesses)

cynthia: it is a design tool discussion. you don't put prompts in peoples' faces
... WYSIWYG tools is the next piece of accessibility work

judy: Cynthia can you keep watching the dialog on this?

<judy> [judy suggests capturing and parking these issues somewhere. any good somewhere to capture and park them?]

JF: I just want to add Cynthia that the meta generator will have the unintended result of not being able to use the validation tools

judy: I want to move to another issue

cynthia: who is working on meta generator and when are they talking?

<judy> We need to recheck this mail http://lists.w3.org/Archives/Public/public-html-a11y/2011Apr/0423.html within this section == On the Co-Chair's decision on the presence of figcaption ==

judy: I am pasting in the next place to look at

<judy> look again please at l http://lists.w3.org/Archives/Public/public-html-a11y/2011Apr/0423.html within this section == On the Co-Chair's decision on the presence of figcaption ==

judy: in the absence of protests we may be heading in the right direction
... I think this one maps against an issue regarding support for native accessibility or legacy issues
... my discussions with people is that their is not a clear position on the task force
... Cynthia, is this a new topic?

cynthia: yes

Stevef: extensive meetings with the waicg, where figcaption is an exception for the case where we agreed you could have a figcaption without an alt

janina: I don't see figcaption in that document we produced.

judy: I recall the legend stuff. In preparation for my response where figcaption is extremely different from an alt.

<oedipus> note: old HTML5 verbiage LEGEND is now HTML5 FIGCAPTION

<judy> [steve clarifies that figcaption was called figlegend at that time]

Stevef: there are many circumstances where you have a graph or a chart where you provide the text alternative. you want to identify there is an image there

<Laura> have to drop off now

Stevef: if you provide the text alternative outside the alt you supply the alt for the label

judy: I do recall the discussion about legends
... I suggest we take this off line
... let's see if we get some agreement.
... on the appropriateness question that is one issue.
... the legacy issue would factor into this as well

<oedipus> example of CAPTION versus ALT versus LONGDESC: http://www.w3.org/html/wg/wiki/LongdescRetention#Gregory_J._Rosmaita.27s_Original_Rationale_for_Retention

cynthia: some of the legacy issues are better handled in HTML 5

judy: what I would like to do is figure out the remaining status on each section
... would you be able to do a joint update on the table summary?

gregory: yes

<Laura> Issue 30 Change Proposal: Include longdesc in HTML5: http://www.w3.org/html/wg/wiki/ChangeProposals/InstateLongdesc

judy: john update on table summary?

JF: yes

judy: the question of splitting out these mails, they may be better handled discretely

<oedipus> stevef, prose comparing longdesc and alt with caption/figcaption available at http://www.w3.org/html/wg/wiki/LongdescRetention#Gregory_J._Rosmaita.27s_Original_Rationale_for_Retention

<JF> +1 to split out

judy: do people have problem splitting these up and taking them on

<oedipus> plus 1 to split

<Leonie_Watson> +1 to splitting


<Laura> bye

Resolution: No objection to splitting

Judy: we need to get the essential issues left
... the question of where the normative alt guidance suggests?

<Leonie_Watson> /me I'll scribe, with the caveat I may be a little late next week.

<oedipus> JB: scribe next week (May 9) leonie watson -- if late, GJR will scribe until LW ready

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2011/05/02 17:10:38 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.135  of Date: 2009/03/02 03:52:20  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/s review//
Succeeded: s/read all of maciej'/read all of maciej's review/
Succeeded: s/specificalloy/specifically/
Succeeded: s/I offered suggestions to help the chairs work with the people./Janina and I also mentioned our reply to Paul that we were hoping to work with the chairs on this./
Succeeded: s/lisa/leif/
Succeeded: s/atl/alt/
Succeeded: i/JB: may have short window/scribenick: oedipus
Succeeded: i/RS: not aware that browser vendors do not want to provide accommodation/scribenick: JF
Succeeded: i/judy: on the second section on meta generator/scribenick: richardschwerdtfe
Succeeded: i/RS: posted text to list -- read all of maciej's review/scribenick: oedipus
Succeeded: i/judy: this is an email from Sam/scribenick: richardschwerdtfe
Found Scribe: rich_schwerdtfeger

WARNING: "Scribe: rich_schwerdtfeger" command found, 
but no lines found matching "<rich_schwerdtfeger> . . . "
Use "ScribeNick: dbooth" (for example) to specify the scribe's IRC nickname.

Found Scribe: Rich
Found ScribeNick: rechardschwerdtfe
WARNING: No scribe lines found matching ScribeNick pattern: <rechardschwerdtfe> ...
Found ScribeNick: richardschwerdtfe
Found ScribeNick: oedipus
Found ScribeNick: richardschwerdtfe
Found ScribeNick: oedipus
Found ScribeNick: JF
Found ScribeNick: oedipus
Found Scribe: rich
Found ScribeNick: richardschwerdtfe
Scribes: rich_schwerdtfeger, Rich
ScribeNicks: rechardschwerdtfe, richardschwerdtfe, oedipus, JF
Default Present: Gregory_Rosmaita, Cynthia_Shelly, Judy, Rich, janina, +1.650.862.aaaa, John_Foliot, Michael_Cooper, Leonie_Watson, Stevef, +1.617.365.aabb, Laura_Carlson, Geoff
Present: Cynthia_Shelly Geoff Gregory_Rosmaita John_Foliot Judy Laura_Carlson Leonie_Watson Michael_Cooper Rich Stevef janina
Regrets: Lynn_Holdsworth Marco_Ranon
Agenda: http://lists.w3.org/Archives/Public/public-html-a11y/2011Apr/0424.html
Got date from IRC log name: 02 May 2011
Guessing minutes URL: http://www.w3.org/2011/05/02-text-minutes.html
People with action items: 

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

[End of scribe.perl diagnostic output]