a11y TF Text sub-team

03 Apr 2012

See also: IRC log


John_Foliot, Judy, David_MacDonald, Janina, Leif, Joshue108, Laura_Carlson, [IPcaller]
JF, , David, Josh


<JF> scribe: JF

additional coordination update on several text-alternatives change proposals, and other pending issues

Janina: have determined a consensus with Objections of latest resolution

2 long-standing objections that were expected, and one comment about use of word "confirm"

word-smithing issue rather than substantive issues

<laura> Hi Josh

Judy: that was more related to Item 2

issue 204: (a) response on aria 1.1 timeline in relation to html5, and (b) specific 204 response clarification needed

Judy: we had pressed the Chairs on what to do to move along issue 30

clarification on timing issue was critical to Chairs

thus janina's processing of the resolution (for Issue 204 - ARIA 1.1 timing)

Judy: with regard to Issue 204

we believed that what we need is a no change CP

but it seems that the HTML5 spec and the ARIA spec are out of alignment

so there is a requirement to re-sync of the 2 specs

Janina: Sam pointed out a section 5.1.2 (impact of @hidden in HTML5) that was out os sync

we are still working on these sections, so this is *hopefully* going to be an easy [sic] process

Janina: hope to get a sense from other groups meeting this week (UAAG, PF, etc.)

<David> scribe:

<David> Scribe: David

<Joshue108> JF: The issue around this is that it all has to do with the way the a11y API is processing stuff thats not on screen. If we move stuff into a <div> offscreen, that can be processed as flat text.

<Joshue108> JF: So if this is off screen, and flat as such you may not get the structural stuff, so we recommend that you don't do it.

<Judy> scribe: Josh

<David> jf has to do with the way accessibility api handles off screen...if we move to off screen... div... but it flattens... can use aria describedby to point offscreenbut it is flattened

<Joshue108> JF: Can you rephrase that Janina?

<Joshue108> JS What I hear is that relating to 204 doesn't talk about various aria described techs not working.


<Joshue108> JF: If you look at that issue.. URL coming now

<Joshue108> JF: Reads issue..

<Joshue108> JF: So I'm frustrated that they don't know what they are asking...

<Joshue108> JF: I'm explaining the issue, lots of hidden content that you cant reference.

<Joshue108> JS: Ok, but our responses on 204 need to focus on aria-hidden etc.

<Joshue108> +q

<Joshue108> JF: Is this aria-hidden or @hidden?

<Joshue108> JS: aria-hidden, this relates to an expemption for aria in 204

<Judy> scribe: David

<scribe> scribe: David

jf the issue with 204 is it says that as an h2 on the page, aria-hidden... aria is out of scope... for html group

<laura> John's CP: http://www.w3.org/html/wg/wiki/ChangeProposals/ARIA_Can_Only_Refer_To_Hidden_Content_With_Specific_Restrictions

janina: Sam is looking for html to say something different from what aria says... that's it... nothing more

jf: i understand that 204 rerers to aria described by offscreen... no distinction... both @hidden, aria-hidden supposed to map to the same accessibility api..

josh... as WCAG-PF ARIa techniques group... we talked about techniques relating to this... for WCAG technique... could john and laura put something together, because you are familiar what you want... we don't know as well what you want...

josh... maybe we're over reaching a bit... originally longdesc... just for blind, now it's extanding, risk being poorly helping everybody

judy: important question.. right now would not spend a lot of effort on details of techniques... need to get it right in the spec first before elaborating on it. applies to 204 and meta etc..

<Joshue108> +1 to Judy

<JF> Scribe: JF

<Joshue108> +q

Janina: We need to stick to simple things to say

Issue 204 is a distraction, but needs to be answered

<janina> David, I think this is described in our minutes from last Tuesday's call

Judy: JF and Janina to take Issue 204 off list for further discussion

may need to find an author to write that, to deal with the process issue

Judy: to a comment from Josh - if people have techniques ideas for HTML5/WCAG please send contributions to that group

additional coordination update on several text-alternatives change proposals, and other pending issues

<Joshue108> Please send tech ideas etc to the html-techs-tf list and/or ping me

Judy: coordination meetings with 2 of the Chairs (1 on vacation), to discuss when things got confusing (re: Process)

met for the past 2 weeks on Mondays (1 hour)

looking at sequence of issues in depth, including @longdesc and the Issue 204 survey

also meta name generator, and how @alt text is dealt with

then <canvas>

outcomes are very specific actions (clarifications, CP's etc.) so that they (the chairs) can process their work queue

this may seem counter intuitive at times, but helps get clarity

<Joshue108> -q

issue 30: expected next longdesc steps

Judy: due to Issue 204, it must be closed before we can take up Issue 30

so next step time-wise is to address 204 as simply as possible, after which Issue 30 will be taken up

so request is to "hang on"

JF: any indication of a date at all?

Judy: is directly tied to Issue 204

the only way out is through...

checking status of other issues pending longdesc resolution: issue 203, possibly 202

Judy: Issue 203 - this is a quick thing (I think) which was missed Monday

but it close(d) midnight Monday

<David> scribe: David

jb: process wise.. we haven't broached idea of dependency delay but is worth talking about (dependent on Issue 30)

jf: can we move forward without an aria solution.... need an accessible name to the video ... traditional was the alt but can't put that on video... but aria provides short description... need a long text image for the poster... would look like described at
... dsome want to put transcript and desciption of the poster together... no native solution in html5... we can do aria label, labeled by... long term onscreen could use described by

<Zakim> janina, you wanted to ask whether we have a non-ARIA alternative for 203?

janina: if we get longdesc back , we couldn't put 2 longdesc on the same element there fore need another attribute

jb: needs to depend on longdesc for now, and then queue this up for aria 1.1 for more evolved version

jf: i couldn't give details on 203 because of dependencies, chairs did not accept that...

<laura> I have to drop off the call. Bye.

<JF> scribe: JF

<Joshue108> I also have to drop off, good call. Bye.

<David> jb: let's follow this up offline... we'll raise the issue of dependency delay

meta-name generator (a) next CP steps; (b) core argument; (c) discussing bugs

Judy: a) meta name generator - want to thank Laura for setting up a wiki for working on that issue

have a bunch of details in "my head" which will be added by this time next week

there were comments w.r.t. weighting of arguments, were as others wer flat wrong

so a new CP will focus on core argument, and focus on the larger issue

some of the assertions of the original CP are unverified themselves

so we need to assert that despite what others might think, you *can* add @alt text inside of a CMS, and the need for sensible @alt text has never gone away

Judy: hope to review 2 open bugs, to see if 1 of the bugs might derail or cause damage

<LeifHSilli> https://www.w3.org/Bugs/Public/show_bug.cgi?id=16572

<LeifHSilli> https://www.w3.org/Bugs/Public/show_bug.cgi?id=16126

better to address now rather than later

Leif: think that both are essentially the same, however the first is/was filed against the validator

the idea is that the validator should be able to notify the author, despite the metagenerator string, alt text is not checked

Judy: on the validator bug, if the problem is that it should be addressed in the valdiator warning that @alt text was not checked, isn't that making an excuse?

if ALT isn't there, it sholdn't conform


David: it would be totally against WCAG 2

can't see how we can endorse something like that

we represent people with disabilities, and this is counter to their needs

Leif: HTML5 already says that if there is a metagenerator, and if alt is not used correctly, then the document is not conforming

The problem is not that the doc isn't non-conforming, it is that the Validator is not issuing a failure notice

David" we want to make it as "Painful" as possible when authors don't provide alt text

Leif: not sure difference between warning and error

what I am proposing is that the validator is weak - it doesn't impact on anything

Judy: the metagenerator shouldn't affect the error message

<David> scribe: David

jf: the problem is... errors are more substantive than warning... errors are a STOP... warnings are not as strong... a validator is a user agent...it's living a pass... theyshould get validator fixed

<JF> Judy: I think I better understand where Leif's comments are coming from

<JF> would like to look at the second bug

jf: appreciate that there is a validator tie in... let's look at the next bug...

<JF> scribe: JF

Second bug sounds like a complete excuse to the author

Leif: if the editor inserts a generator string, then the validator must indicate that @alt conformance was not checked

I don't completely agree with Judy's perspective

David: want to refocus on why we exist as a group, to make the web more accessible

others can argue against accessibility, we want to make sure that conformance checkers make things accessible

if a validator skips a chunk of code, webmasters will not go check

it is unhelpful to creating a good culture

Judy: if we are attempting to fix the core issue, why present "fallback" ideas?

meta name generator shouldn't stop a conformance checker, and so suggesting that if it does, it should warn the author presumes that a conformance checker might not be checking for @alt

which is what is broken

Leif: do you plan to say that if a Validator does not warn the author then nothing be said?

Judy: we could provide advice on what the Validator messages would be appropriate

David: we seem to be arguing against ourselves with these 2 bugs

Leif: this is a compromise proposal if we don't succeed

Judy: we cannot accept for a compromise here

this is the needs of specific users

<janina> I'm sorry, I need to go now

David: concern that we are having a discordant voice here

Leif: I will need to reflect on this, I did not realize it was such a problem

<LeifHSilli> I have withdrawn my bygs.

Judy: thanks all, we didn't get to everything, but we are scheduled to meet next week

rrsagent: Make minutes

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.136 (CVS log)
$Date: 2012/04/03 18:38:45 $

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/same/Sam/
Succeeded: s/th eonly/the only/
Succeeded: s/jb/jf/
Succeeded: s/depend on longdesc/depend on longdesc for now, and then queue this up for aria 1.1 for more evolved version/
Succeeded: s/longdewc/longdesc/
Succeeded: s/jb/jf/
Found Scribe: JF
Inferring ScribeNick: JF
Found Scribe: 
Found Scribe: David
Found Scribe: Josh
Found Scribe: David
Inferring ScribeNick: David
Found Scribe: David
Inferring ScribeNick: David
Found Scribe: JF
Inferring ScribeNick: JF
Found Scribe: David
Inferring ScribeNick: David
Found Scribe: JF
Inferring ScribeNick: JF
Found Scribe: David
Inferring ScribeNick: David
Found Scribe: JF
Inferring ScribeNick: JF
Scribes: JF, , David, Josh
ScribeNicks: JF, David
Default Present: John_Foliot, Judy, David_MacDonald, Janina, Leif, Joshue108, Laura_Carlson, [IPcaller]
Present: John_Foliot Judy David_MacDonald Janina Leif Joshue108 Laura_Carlson [IPcaller]
Got date from IRC log name: 03 Apr 2012
Guessing minutes URL: http://www.w3.org/2012/04/03-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]