HTML-A11Y Text Alternatives Sub-Group Teleconference -- 25 Apr 2011

JB: 3 new ones - at least 2 fall under this subgroup's purview; third might as well
... one has to do with validation of @alt
... another normative guidance for @alt
... validation of @alt -- Rich and i began to mock-up a draft of something to review in response -- consult http://lists.w3.org/Archives/Public/public-html/2011Apr/0451.html

<judy> http://lists.w3.org/Archives/Public/public-html/2011Apr/0451.html

JB: 6-part decision -- 6 sub-items on whether HTML5 validates with or without presence of @alt, @title, FIGCAPTION, etc.
... response missing info on how @alt works as opposed to @title (@alt has default place in visual rendering; @title does not and is transitory
... advice on Alt Text Techs -- WAI CG has interest in responding to this

http://www.w3.org/WAI/CG/

ISSUE-31 / ISSUE-80 requirements survey: http://lists.w3.org/Archives/Public/public-html/2011Apr/0453.html

JB: what text to use for IMG element definition

GJR: this is our 1 win
... changed precisely in way submitted to HTML WG

JB: further action needed?

<JF> +1

<janina> +1

SF: nothing further

plus 1

<MRanon> +1

JB: other items fall in scope of this group?

SF: yes

GJR: yes

JF: yes

<janina> yes

JB: any objections?

agendum 3

JB: quick check for now

<judy> http://www.w3.org/WAI/PF/HTML/wiki/Verbose_desc_reqs#Requirements

JB: discussed last week whether having agreed-upon set of reqs from us would be helpful
... 9 items currently
... Rich only person to provide comments since last week
... anyone else have chance to review Verbose Desc Reqs this week and thoughts upon them?
... thanks to RichS for comments -- GJR integrated some and documented others

http://www.w3.org/WAI/PF/HTML/wiki/Talk:Verbose_desc_reqs

JF: will look at it this week

JB: any comments?

JF: clarification -- further feedback, straight up review -- specific deliverable?

JB: specific requirements may help form consensus on this TF for clarification -- haven't walked through too many clarifications as consensus activity, some of these may map against arguements/lack-of-agreement
... hoping to come up with good consensus capture of overall principles and figure out points-of-discussoin -- like to address this week on list and next week at call
... like to spend more time talking about clarification emails today
... will review requirements

JS: will review

reviewers: JF, JS, JB (already reviewed RS, GJR)

JF: request if comment use the [text] subject line tag

GF: will review, too

Draft clarification on title/alt (Rich/Judy drafting, will

JB: @alt and @title validation -- sent email to RS this morning -- can you clean-up and bounce-to-list?

RS: send note to list?

JB: suggested that add edits i made, strip out what indicated, and post to list -- can do myself

RS: please do JB

JB: posting now
... Rich and i looked at the chairs' decision on validation of @alt which has 6-sub-positions in it
... noted that there appear to be some things that chairs' didn't understand as reflected in decision; collection of decisions of considerable concern; took premise that while a FO may need to be forwarded, wanted to reply to decision specifically
... 1 approach: respond to everything incorrect in decision, or highlight most important mistakes/errors
... RS found that 4 of 6 sub-decisions problematic
... would like to know if attendees agree with conclusion, and that this email captures subgroup's understanding
... 4 items: 1) aria-labelledby does not make @alt conforming; 2) role="presentation" does not make missing @alt conforming; 3) missing @title ok if no @alt; 4) FIGCAPTION

<JF> +Q

JF: skipped over meta name="generator"
... if put meta name="generator" in HEAD would allow author to not add any @alt AND validate
... personal email exception -- i shape my email in accordance with the person to whom i am sending the emessage

JB: any disagreement that need consensus clarification on 1) aria-labelledby does not make @alt conforming; 2) role="presentation" does not make missing @alt conforming; 3) missing @title ok if no @alt; 4) FIGCAPTION, 5) meta name="generator"

SF: figcaption issue?

JB: 1st reaction, caption can't stand in for @alt
... looked at material on-line -- FIGCAPTION in publishing has specific purpose with nothing to do with @alt -- haven't had chance to check against HTML5 draft; mis-match of purpose in my opinion
... sceintific publication, have terse caption that encapsulates image context, but not sufficient as @alt

<JF> +q

GF: agree with JB -- FIGCAPTION used for totally diff purpose than @alt -- not sure if strictly used as visible label, but conflating the 2 is a HUGE mistake

SF: allowing use of FIGCAPTION not to replace @alt --
... @title becomes caption below image -- if person can't provide @alt, if do provide CAPTION for it, will be conforming, but not neccessarily accessible
... cases where users can't or will not provide @alt

JF: this is a problem, but this is the least of the issues facing us
... if i post pic of cat on flickr and use caption "the neighborhood cat" --- need to investigate positive implications
... is caption appropriate @alt text? better than 73525.jpg

JB: like to review with Geoff -- FIGCAPTION use generally and specifically -- may be substantially different
... like some examples
... goal of what is conforming is something that is accessible -- whatever we agree to in TF, has to be something specific

GF: will work with Judy on this

<judy> ACTION: judy , geoff to look into figcaption & alt decision [recorded in http://www.w3.org/2011/04/25-text-minutes.html#action01]

JB: anyone have questions about other items RS and i identified?

JS: a bit confused -- don't want to set policy on bad UI design (flickr case)
... concern about conflating caption and @alt
... caption like a comment on IMG; @alt is description of IMG

RS: aria-labelledby conforming -- rationale -- if author uses aria-labelledby to point to visual image, when turn image off, will have label present with image, so operates same as @alt
... use @alt or @aria-labelleby -- with labelledby saying this belongs to this particular image -- label probably centered in area reserved for image
... similar to figcaption
... showing label associated with image

SF: issues with labelledby is there is no need for text to be physically associated with image -- can be anywhere on page -- when image disappears, could be problematic -- FIGCAPTION has to be inside figure next to image

<JF> +1 to Stevef

SF: with labelledby need text alternative, can't be caption because of way mapped in a11y APIs -- no way to say this is not a text equivalent but a caption -- FIGCAPTION has semantic meaning; can't mistake caption for @alt

RS: view CAPTION as label

JF: more direct association
... looking at web page with list of speakers at conference -- have phone numbers -- if images turned off, have big blank square and loss of binding

<Zakim> janina, you wanted to say Flicker's inadequate ml shouldn't define good enough alt

MR: people use text with images using HTML4 -- use text as label for image, but not programmtically associated -- with HTML5 using CAPTION can be programmatically associated, but htere are cases where caption provides more info than contained in image
... authors can use labels in diff ways -- should provide authoring advice for labelledby and FIGCAPTION
... as we did for @alt

JB: appreciate discussion --
... wonder if people could take on individual sections to refine and post to list in next few days so can file comprehensive clarification email request by friday so can look at it at next monday's meeting and vote on consensu

<JF> +q

JB: level of detail -- would like to explore specific coordinated comprehensive clarification on each item and then sending them to HTML WG chairs with some urgency

JF: concerned about meta name="generator" -- currently discussion on list
... Leif raised some really good points and done very good research about auto-generated meta strings

JB: would like to proceed as quickly as possible -- terse response useful, may want to hold 1 or more aside to get full consensus -- idally would be good to pass along whole package at once

starter draft response (JB and RS) http://lists.w3.org/Archives/Public/public-html-a11y/2011Apr/0287.html

JB: would generator change to decision or clarification presented

JF: if start multiple attacks on individual issues, may be subject to divide and conquer counter-strategy
... will chairs accept sub-decision comments individually?

JB: invite discussion --- suggest that we do as much work as possible, so reply to as much as can now, perhaps add comments later
... with meta name="generator" sounds like JF willing to help with drafting?

JF: yes

JB: Rich do you want to clarify the aria-labelledby sub-decision

RS: don't need to argue with decision -- don't care if made argument or not -- using aria-labelledby is not overwritten by @alt

JB: drop aria-labelledby?

[no objections]

RS: can still use, NOT a replacement for @alt

JB: support clarification as written?

JF: preence of role="presentation" should not make @alt non-conforming

RS: failure condition if have @alt and role="presentation"

JS: think they want alt="" for presentational images

RS: redundant

JS: agree

JF: authoring tools will always insert alt="something" -- if don't put in value, most will put in alt="" -- presentation role is additional info -- whay if add one remove other?

RS: if marked role="presentation" no reason to add alt="" -- author designated as presentational

JS: problem other way around -- encourage use of role="presentation" less ambiguous than alt=""

<JF> not a hill I'm willing to die on

RS: @alt with role="presenetation" eliminates need for alt="" and includes it in A11y API level -- want to keep presentation from a11y APIs -- stuck with @alt

SF: agree with RS, but in HMTL5 says @alt="" is same as role="presentation" which means that any img with alt="" is equivalent to role="presentation"

RS: either or correct?

SF: prefer to use role="presentaion" because is clearer semantically -- counsel use both or one (role="presentation")

RS: role="presentation" does what we need

SF: role="presentation" is in a11y layer; alt="" will be represented differently in view where images disabled -- if role="presentation" won't treat same way -- need to treat null alt as role="presentation"

JB: could RS and SF take this discussion to email and report back to the group?

SF: yes

RS: will do my best

JB: rescanning 6 issues: aria-labelleby decision ok?
... role="presentation" needs more info from RS and SF

<scribe> ACTION: Rich and Steve to draft reply to role="presentation" sub-decision for discussion at next week's meeting [recorded in http://www.w3.org/2011/04/25-text-minutes.html#action02]

JB: please review contents of http://lists.w3.org/Archives/Public/public-html-a11y/2011Apr/0287.html and let us know if section beginning "title"...

"Unlike alt="", role="presentation" has the added value of removing the image from the accessibility API object tree, effectively filtering out the image and improving assistive technology performance. Furthermore, a role of resentation is to state the intent of the author in a declarative fashion. For these reasons, role="presentation" should be considered a suitable alternative to...

scribe: requiring alt when it adds no meaningful information to an AT."

SF: agree with what is in email, have further comments and ideas
... will list what i perceive as problems, add to RS and JB's prose and repost to list

JB: propose your terse addition? link to more detailed explanations fine

SF: just add terse recommendation with links as needed

<scribe> ACTION: Steve - add terse statement about role="presentation" to RS and JB's prose and repost to list [recorded in http://www.w3.org/2011/04/25-text-minutes.html#action03]

JB: Geoff -- please respond to SteveF's posts from the past several days

GF: will do

JB: FIGCAPTION needs more investigation and a report back to group

<JF> +q

<gfreed> +q

JB: for FIGCAPTION, would like for us to consense a comprehensive reply to this by next monday's meeting or early in the week, and send these as clarification to the chairs; then see if reclarification is needed; and proceed with formal objections with expedited appeal as needed

JF: when look at 6, 2 critical ones are meta name="geneartor" and @title as replacement for @alt
... FIGCAPTION and labelledby worth looking at but not "dying" for
... severity: @tltle and generator most severe

JB: looking at comprehensive clarification on what we do not agree with in decsions

JF: getting clarification back may be useful

JB: want to get comprehensive clarification request out as soon as possible

JF: suggesting that as move forward, some things more critical to others

GF: agree with JF -- @title in place of @alt is a SERIOUS problem

<JF> +1 t Geoff

<judy> s/early in the week/early in the week, and send these as clarification to the chairs; then see if reclarification is needed; and proceed with formal objections with expedited appeal as needed

GF: don't want to break implementations -- drove home to everyone to use @alt -- changing that to say @title is ok is going to mess up a lot of work already done-- not a good idea period

JB: draft email has very terse clarifications -- appears to me there are multiple misunderstandings in charis' decision
... may be important from POV of priciples

http://lists.w3.org/Archives/Public/public-html-a11y/2011Apr/0287.html

<gfreed> +q

GF: if not mistaken, most SRs come with there presets defaulting to @alt not @title
... SR users often don't personalize settings

JB: add as comment to section?

GF: ok

JB: want to make 3 comments on other decisions -- location of @alt techniques (WAI CG responding to that -- may be able to review in detail next monday

Update on formal objection on normative accessibility

JB: any objections to waiting for WAI CG report/draft
... already a formal objection from SF on this
... clarification that may set basis for formal objection

Continue planning clarification mails on rejected

JB: scope includes rejections on @longdesc, @summary for TABLE, and @poster
... email discussion on each of these -- some very lengthy -- can we prepare parrallell clarification emails
... for @longdesc there is enough material to fill a book -- posibliity of starting more formal dialouge based on terse extraction from @longdesc materails
... poster issue may be easiest to tackle -- JF work with someone to turn into parrallell comment / basis for future formal objection

JF: filed FO on alt poster -- said technical stuff inconsistent, even though requested assistance on technical stuff

JB: would you work with someone (probably JB) to draft next-round clarification and re-draft reply with JB to prepare something for the group to review next monday

JF: sean hayes of MS has offered to help me with technical portion of FO

<gfreed> geoff has to run.

JB: anyone who wants to write a sentence or 2 on alt poster?
... can we have draft clarification email for monday for voting on monday by group
... assume that people have read pertinent emails

<judy> ACTION: JohnF, judy, sean work on reclarification email on poster-alt (alt-poster) [recorded in http://www.w3.org/2011/04/25-text-minutes.html#action04]

JB: techincally "poster alt" but should be "alt poster"
... table sujmmary -- draft of clarification email?

GJR: have a CP for summary as element

JB: GJR can you draft email in format of email RS and JB circulated

GJR: yes, will ping if necessary

<scribe> ACTION: clarification email for @summary for HTML WG chairs [recorded in http://www.w3.org/2011/04/25-text-minutes.html#action05]

JB: @longdesc
... LauraC been in touch -- was going to try to attend last part of call -- may need to try to stablize discussion and get clarification of things at this time that could send from this subgroup to chairs, see what chairs reply and depending on circumstances draft an FO
... may make sense to work on other responses this week to get template and basis for future work

Recap of action items and timelines

RS: Steve and i will look at section on role="presentation" and @alt

JB: Rich ok to scribe next week?

RS: yes

JF: looking at meta generator to produce terse text; working with JB on alt poster

GJR: @summary for table

JB: scribe volunteer for 2 weeks from today?

MR: won't be on call next week (bank holiday in UK)
... won't be available for next 2 weeks

JB: meeting next monday, same time, same IRC channel

[ADJOURNED]

- DRAFT -

HTML-A11Y Text Alternatives Sub-Group Teleconference

25 Apr 2011

Attendees

Contents

Action Item Review

Draft clarification on title/alt (Rich/Judy drafting, will

Update on formal objection on normative accessibility

Continue planning clarification mails on rejected

Recap of action items and timelines

Summary of Action Items

Scribe.perl diagnostic output