Accessibility Guidelines Working Group Teleconference

05 Mar 2019


AWK, Rachael, Chuck, alastairc, Sonja, Laura, Makoto, JakeAbma, Detlev, MarcJohlic, MichaelC, bruce_bailey, stevelee, Brooks, kirkwood, JF, Wilco, Katie_Haritos-Shea, SteveRepsher, mbgower
Glenda, Rafal
Tiffany, Twalters, Rachael


<AWK> Zakim clear agenda

<twalters> Hi everyone.

<Detlev> I can do it

<Detlev> I step back

<AWK> Scribe: Tiffany

<laura> Scribing Commands and Related Info: https://www.w3.org/WAI/GL/wiki/Scribing_Commands_and_Related_Info

<AWK> Scribe: Twalters


AWK: wcag 2.2 CFC

We're moving forward with WCAG 2.2

<kirkwood> +1

<kirkwood> ;)

starting work on requirements for 2.2. More details will be shared at face to face next week.

ACT Survey (https://www.w3.org/2002/09/wbs/35422/ACT-RF_CR/results)

Issue Survey https://www.w3.org/2002/09/wbs/35422/technique-approvals8/results (#4 only)

Approval pending on Accessibility Testing Conformance Rules. Please review and respond to survey.

<bruce_bailey> http://www.w3.org/2002/09/wbs/35422/ACT-RF_CR

<alastairc> https://www.w3.org/2002/09/wbs/35422/ACT-RF_CR/

<bruce_bailey> Links from survey are in introduction section

<alastairc> the people who filled it in all approved with no comments.

<bruce_bailey> links are not available in results view of survey

<bruce_bailey> I though CSUN was first on agenda?

AWK: There are more issues to be reviewed. Got through a few of them last week. How should we proceed?

MichaelC - I've added a draft technique with examples. Initially it was an understanding document.

MikeC - how many people have had a chance to review. Will take feedback and incorporate into next draft.

MikeC - Close proximity comments have been striked out.

<bruce_bailey> Thanks @MikeG: i have not reviewed since "close proximity" was changed to "adjacent"

Understanding document had a restriction label so description had to be in close proximity. May have been hard for others to follow.

<bruce_bailey> s/Thanks @MikeC/MikeG

having a label adjacent to the field is good but not the only technique.

<AWK> Preview version: https://cdn.staticaly.com/gh/w3c/wcag/LabelinNameUnderstanding/understanding/21/label-in-name.html

MikeGower: speech input ...

thanks Alastair.

David Macdonald... something is missing from label in name

<AWK> https://github.com/w3c/wcag/pull/639/files

<jon_avila> I agree with Andrew-- that's the content that matches

<AWK> The intent of this Success Criterion (SC) is to make information conveyed by text that is <em>visually</em> associated with a component match the information that is associated with the component <em>programmatically</em>

thanks bruce!

Chuck: reads this as programmtical and visual matches

<chuck> reads this as programmtical and visual matches

<kirkwood> “vsiually associated” “text incorporated in” associated is difficult to understand

<alastairc> Cache bypassed version: https://cdn.staticaly.com/gh/w3c/wcag/LabelinNameUnderstanding/understanding/21/label-in-name.html?x=2

<bruce_bailey> Yes, the minor edit I suggested has been incorporated. Thank you MikeG.

Regarding legend shouldn't be included in label.

<AWK> "Other text displayed on the screen that is correctly coded to meet 1.3.1: Info and Relationships is not normally factored into the calculation for the accessible name of a UI component without author intervention (via ARIA labeling techniques). The most common of these are:"

the idea is if there is a radio group with options, "do you want to go to lunch" shouldn't be apart of accessibility, but can be included by ARIA. Will be reworded in document to clarify the intent.

<Zakim> bruce_bailey, you wanted to mention that math best practices is problematic

Brooks, we don't want content authors to include group names. good example, to not include legend if or group names because their doubling feedback that impacts customer experience.

<bruce_bailey> Doing so would mean that a user who said "eleven multiplied by 3 is equivalent to thirty-three" would be unlikely to match.

<jon_avila> I agree with the exception for the group bit -- except in the case where the group name is the only visual label such as for phone number.

<alastairc> https://design-system.service.gov.uk/get-started/labels-legends-headings/label-h1/index.html

<Zakim> AWK, you wanted to suggest adding mike's "do you want to go to lunch" radio group example as an example

AWK, confining the label to adjacent text. Is there a better phrase to express this description?

bruce, its a heading not a content. several interpretations of labels to solve for. following the heading as default view is helpful. there are coders trying to meet this because there's no guidance right now.

Bruce, identifying label text seems to be the goal.

jesse, question: left of controls... where should descriptions go?

AWK: trying to determine how strongly should the guidance be worded.

jesse, concerned it will be interpreted as requirement. whether description should be left, right, centered. has approriate relationships.

bruce, all referenced are ok. just need to clarify acceptable methods that may need to be discouraged.

<bruce_bailey> Can someone post github link to make comments?

mikegower, developer should be enforcing visual design.

<bruce_bailey> this is NOT it: http://github.com/w3c/wcag/pull/614

bruce, mathematics advise isn't correct. let's drop for now. example: math ml well supported can still spur issues.

plain text is clearer

<bruce_bailey> Doing so would mean that a user who said "eleven multiplied by 3 is equivalent to thirty-three" would be unlikely to match.

the challenge is every screenreader/at interprets differently.

bruce_bailey, can change the math multiplication symbol.

"x" vs. actual symbol creates confusion

Brooks, what's the label for page control. stand alone icon image as control: when placed on page. where does descriptions go for stand alones?

are we creating accidental descriptions next to that icon.

*thanks Rachael

<Rachael> scribe: Rachael

<Detlev> In that case, users would likely not group the text and the (i) semantically, even if adjacent

The third bullet is trying to act as a button. Its when there is not text label inside the button and people start searching around the button to find the label. In the matching technique, the procedure is text inside the button.

Brooks: I was specifically thinking of the icon where text is associated it.

<twalters> *thanks everyone for jumping in to help me on my first scribe run!

<twalters> *whew

mgower: I will try to get icons worked on.

Brooks: I am seeing this more and more.

mgower: I will try to tackle this. I see the history of the question.

AWK: There are a lot of buttons that have text. Thinking about MSWord or other word processing tools where there is a B for bold or I for italics. Labelling it B isn't really helpful. We have clarification where text is used symbolically.

detlev: I think grouping them semantically, the text next to an i wouldn't be a label. People would know its related to something but its not technically a label.

Brooks: I agree with you.

AWK: Question from Michael Cooper. With a button and input elements, do we count an image?

In punctuation section, there are two controls. Is the potential confusion for the user offset because they are controls? Does that make sense?

Mgower: I need an example

AWK: Check out the example. Mike will tackle the edits. I think we've gone through the comments on this. Is there anything new?
... any additional comments?
... I think the changes Mike will be doing are editorial. Can we accept this as amended or do we need to make it pending something?
... Bruce is not here but he sounded happy with the changes to the math section.

<Detlev> +1

<Chuck> +1

<alastairc> +1 I think it's a good update, can update (again) later.

AWK: Any objection to accepting this as ammended pending changes described on the call?


<laura> +1

<JakeAbma> +1

<Brooks> +1

<Ryladog> +1

<JF> +1

<MarcJohlic> +1

RESOLUTION: Accept as amended pending changes described on the call

<Chuck> +1 keeping Mike's life hostage.

AWK: Will make some changes and circle back with Mike.
... Detlev had one item that wasn't added to the agenda around one of our recent responses. Detlev, can you speak on your concerns?

Detlev: There are several things so which one are you referring to?



<AWK> https://github.com/w3c/wcag/issues/80#issuecomment-469206007

Detlev: I started a discussion about background images used for controls. I'm not sure what the original intent was but even if the user picks their own colors then the item will disappear visually. Its only related to text alternative 1.1.1. We often have cases in testing where people use background images as a control and add aria-label.

This supports screenreaders but it causes problems for users who use their own stylesheets or adjust in browser. We discussed this in 2015 and we were going to amend the failure to be more specific to text alternatives so not apply to disappearing background images.

I think we need to go back to this and create another failure for images disappearing when users apply their own styles.

<alastairc> Might need to remove the note.

AWK: So you are not suggesting that the existing failure is incorrect but that we need an additional failure?

Detlev: There is an issue with the existing failure that the "label is available" but it only addresses available to screenreaders vs available visually to users with custom style sheets.

<AWK> https://www.w3.org/TR/WCAG20-TECHS/F3.html

AWK: link to current technique F3
... Procedure Examine all images added to the content via CSS, HTML style attributes, or dynamically in script as background images. Check that the images do not convey important information. If an image does convey important information, the information is provided to assistive technologies and is also available when the CSS image is not displayed.
... So I'm not sure where the label comes in.

Detlev: Perhaps it has been changed. When I looked at this last it was still ambiguous but there was a pull request associated with it. It has not been merged.
... There is still concerns with wording.

AWK: This will take some research to discover why it wasn't changed unless someone else remembers this discussion from 2016 Q1.

Detlev: I think there was a request not to apply this change until there was a new failure but that failure was never created so this has been shelved.

<alastairc> Can see an argument that it needs focusing on the programatic aspect, separate one for disappearing background (maybe), but the UAs are updated.

This creates a problem for us when dictating failures. Is it a failure of 1.1.1 and if not, where does it belong?

AWK: I will look into it and see what happened and then we can revisit it.

RESOLUTION: Andrew will investigate Pull Request 120

<Chuck> If we are in free-form, I'd like to discuss: https://github.com/w3c/wcag/issues/542

AWK: So lets take some time to look at new issues that need someone assigned to them. I encourage people to sign up for some of these.

<AWK> https://github.com/w3c/wcag/issues

For our CSUN face to face so if you have issues that you can make a pull request, we'll have a chance to review some and conclude some of them. Its not a major focus because we have the charter and gearing up for 2.2 but we do have some time set aside....

On the list of issues, there is one on reflow. There is one about testing with pointer gestures. One about testing orientation without a mobile device. There is a question we may be able to address on the call, if a field asks for information about your organization is that covered under SC 1.3.5?

It says, if the field says the zip code for "your organization" does that apply? Jon Avila, are you on the call?

AWK: If a field on the form says "your organization's zip code" is that covered by 1.3.5? My take is no.

JF: Agree with that.

AWK: Does anyone feel differently about that?

<alastairc> https://github.com/w3c/wcag/issues/642

AWK: Organization is one. My name, my last name, my organization is covered but when it dives deeper to my organization's zip do we agree that is not covered?

detlev: My organization is covered but I'm unsure about the level deeper.

JF: If we think about it from the perspective of the technique, what is the information likely captured by the browser? It's scoped to inputs related to the user. I'd have to go check what the actual tag. I don't know if there is a tag called "my organization"

<alastairc> https://www.w3.org/TR/WCAG21/#input-purposes

AWK: What would that apply to then? Would it be me an employee vs me as the owner?

JF: The taxonomic term is organization "Company name corresponding... with other information associated with field." I take back my original comment.

AWK: So Organization is covered. Is the organization's zip code covered?

JF: postal code? I would assume that the postal code that is stored would be the person's address not the organization's address. So first name: Andrew, last name: Kirkpatrick, Organization: Adobe, Zip: home zip code.

So the zip code would be the person's not the organization's.

AWK: You and I agree. Alastair?

Alastairc: I think this falls into the category of lets be conservative. The information we have is a bit vague but from the point of view of the user sitting there, it can get complicated. I have separate browser profiles. My work computer has my work information and my home computer has my home information.

JF: And some browsers allow you to store multiple profiles.

Alastairc: If we keep the simple situation, then everything applies to the person. We are not trying to require the organization information, but like our previous decision to not fail fields that use the tags, we likely shouldn't fail this field for using the autocomplete tag.

JF: I think we need to approach it as a technique or add it to the understanding document, I think we need to include some patterns of what we expect to see.

MichaelC: I want to point out that the content is pulled from HTML5. I don't think we should interpret it without working with the HTML Working group.

JF: The nuance we added is that it is only information about the user that requires it for accessibility.
... we didn't pull over the autofill attribute grouping constructs. We also have somewhat have a grouping function for forms that most people use only with radio buttons and checkboxes but people could use grouping for collecting personal information.

AWK: I will write a proposed response.

<AWK> "Proposed WG Response: 1.3.5 Identify Purpose is designed to apply to input fields about the user and this includes some organizational information (specifically Title and Organization name) that applies directly to the user. Additional information about the organization such as the Postal code for the organization does not apply specifically to the user and is not included."

Chuck: Is your proposed response going to be for or against? I lost track of where people stand on this issue.

AWK: answer pasted above.

Chuck: I would argue against it. Some browsers can track multiple profiles. Where it can be programatically determined what the person's work zip code is, we should capture it.

Alastairc: I don't think we have the granularity.

Chuck: not now but we may in the near future.

AWK: I think the question is "must" or "should"

Chuck: It must not be "must"

AWK: I will still work on the response. Any other questions around this one?
... Any other open new issues?

Chuck: If we are free form, can we discuss 542?

AWK: We have a few other new issues. No one is assigned to any I've listed. Please sign up.
... Next question: Will anyone have anything ready to review to get on survey?

<Detlev> I had the monkey technique for single key shortcuts

Rachael: I will have the 3 motion actuation techniques ready for survey.

<Chuck> https://github.com/w3c/wcag/issues/542


Chuck: I assigned this and then came to understand that I have a conflict of interest due to company. Can we unassign me and reassign me?

AWK: The question is around duplicate attributes causing a conformance error around SC 4.1.1.

<twalters> is there a specific toolkit the wg uses to test these issues?

Alastair: There is a lot of information in this. The more I dug in, the harder it was to justify parts of the SC as things that are not covered elsewhere. I've struggled to find some solution that doesnt' end with deprecating the SC

If anyone feels strongly about SC 4.1.1 jumping in to justify it would be good.

<twalters> *ok. so just an automated and manual approach is sufficient.

David: I think we may be able to map the SC and explore this.

Alastair: If major orgs are working around this, does it undermine the AG purpose?

David: My experience is that we don't test this. It would be difficult to go to court and prosecute this failure.

Brooks: This is where we get into an area where we don't control or coordinate with the tools that create the user experience. The default is to put the onus on the content owner. This is a perfect example of needing to approach the entire ecosphere (AT, content, etc).
... there is some reason to what I've been recommending about needing to navigate to a different environment that takes more into account.

Alastair: When it was put into place, I think it was such an unknown what AT would do but the browsers and AT are better specified and more consistent.

JF: David was at one point asking me. AxeCore does call this as a failure. I agree that part of the issue with this is addressing how browsers handle it. It would be nice to deprecate this.

I think that WCAG 2.0 was specific about what was needed.

AWK: Is someone willing to work on this?

JF: Should we ping the taskforce to get there thoughts? I can talk to Wilco directly. Due diligence on this would be to get their thoughts.

AWK: We can send out a note to them?

<mbgower> IBM's DAP tool also gives this as a failure. As a canary in a coalmine rule, if the dev resolves duplicate IDs, it removes those as a potential cause of an a11y issue

JF: Can we deprecate this in 2.2? I don't know what that means for backward compatibility? I think its benign but we should also check with regulatory people to see what the effect is.

Alastair: If we deprecate it, we would not be able to say anymore that if you meet 2.2 you meet 2.0.

JF: We could say that this area is now covered at the browser level. Will this make any difficulty for organizations about adopting 2.2?

Alastair: Apparently there is this conference next week where we might be able to ask that question.

JF: I can follow up with Lainey Feingold. Can we follow up with Bruce? I just don't know but I think we should ask.

Katie: Right now the US is not requiring 2.1. It might be a problem if they want to use 2.2 but I don't have the answer to it.

JF: The UK and EU would also be good to check with.

Alastair: We currently are under the EU but Shadi would also be good to ask.

<Brooks> The inability to deprecate or otherwise material change any of what's come before in WCAG 2.0 and WCAG 2.1, is one of the big drawbacks, in my opinion, of moving to WCAG 2.2.

<AWK> AWK sent email to ACT TF co-facilitators

AWK: I think we're in good shape for today. We will be sending out a link to an update to the silver requirements which will be the first topic on Monday with the Silver task force. Read it and be prepared to provide comments. Where there are strengths and weaknesses in this draft.
... The other is the ACT survey. Please look at that one as well. We will send out additional details on the agenda in the next day or two.

Does anyone have additional items or questions?

Katie: Did we decide on a dinner on Monday night?

AWK: I don't think so but its a good idea.

Katie: I will take a shot at it. How many are we expecting?

AWK: 15. There are also people from Silver who have not RSVPd on ours. My guess is 22.

JF: There are multiple hotels clustered. My recollection is that there are not a lot of good eating options in that area but if you walk to the disney area, there are mulitple restaurants there that you don't have to pay to enter. Its about a 5 minute walk.

AWK: Thank you.

trackbot end meeting

Summary of Action Items

Summary of Resolutions

  1. Accept as amended pending changes described on the call
  2. Andrew will investigate Pull Request 120
[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.154 (CVS log)
$Date: 2019/03/05 17:57:24 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.154  of Date: 2018/09/25 16:35:56  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

Succeeded: s/tiphanie.walters@wellsfargo.com//
Succeeded: s/zakim: wcag/AWK: wcag/
Succeeded: s/Bruce_Bailey - There/AWK: There/
Succeeded: s/MikeC -/MikeG:/
FAILED: s/Thanks @MikeC/MikeG/
Succeeded: s/MikeBowers -/MikeGower:/
Succeeded: s/Chuck, /Chuck:/
Succeeded: s/Bruce, /AWK:/
Succeeded: s/monke/monkey/
Default Present: AWK, Rachael, Chuck, alastairc, Sonja, Laura, Makoto, JakeAbma, Detlev, MarcJohlic, MichaelC, bruce_bailey, stevelee, Brooks, kirkwood, JF, Wilco, Katie_Haritos-Shea, SteveRepsher, mbgower

WARNING: Replacing previous Present list. (Old list: AWK, alastairc, shadi_, JakeAbma, stevelee, MichaelC, Brooks, Laura, Lauriat, maryjom, kirkwood, Bruce_Bailey, Raf, KimD, JF, david-macdonald, SteveRepsher)
Use 'Present+ ... ' if you meant to add people without replacing the list,
such as: <dbooth> Present+ AWK

Present: AWK Rachael Chuck alastairc Sonja Laura Makoto JakeAbma Detlev MarcJohlic MichaelC bruce_bailey stevelee Brooks kirkwood JF Wilco Katie_Haritos-Shea SteveRepsher mbgower
Regrets: Glenda Rafal
Found Scribe: Tiffany
Found Scribe: Twalters
Inferring ScribeNick: twalters
Found Scribe: Rachael
Inferring ScribeNick: Rachael
Scribes: Tiffany, Twalters, Rachael
ScribeNicks: twalters, Rachael

WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Found Date: 05 Mar 2019
People with action items: 

WARNING: IRC log location not specified!  (You can ignore this 
warning if you do not want the generated minutes to contain 
a link to the original IRC log.)

[End of scribe.perl diagnostic output]