W3C

- DRAFT -

Accessibility Conformance Testing Teleconference

06 Apr 2023

Attendees

Present
Helen, ChrisLoiselle, Wilco, trevor, thbrunet, kathy, present, catherine_droege
Regrets
Chair
SV_MEETING_CHAIR
Scribe
thbrunet

Contents


<catherine_droege> +present

scribe+

ACT Standup

Wilco: Sent out some CFC's. Worked on templating for secondary requirements. ARIA defined attributes PR merged, ready for CFC

Catherine: Nothing to report

Kathy: Prepared for label in name. Looked at items working on for Github guidance. Will have something for that group to look at soon

Chris: Working with Helen tomorrow on the manual test projects. Can take action on Kathy's Github stuff if needed.

Trevor: Made some example test cases that we can look at today as part of state conversation.
... For aria state of properties permitted. Split the expectations one two rules. Can I include these both in on PR?

Wilco: I had a similar situation. Splitting them may advance faster, but you can keep them together.

Trevor: I'll see what comments are and can split if needed.

Chris: I have Lori Oakley that I wanted to introduce as a listener today.

Tom: No report

Helen: Meeting tomorrow with Chris.

Lori: I am employee of Oracle. Division accessibility lead for CX Service. I review and test and coach for our division, which has ~898 people. Can be a bit challenging, but I find the intracacies of WCAG interesting. I thought this might be something I'd be interested in participating in.

Wilco: Welcome

Secondary requirements

<Wilco> https://deploy-preview-200--wai-wcag-act-rules.netlify.app/standards-guidelines/act/rules/bisz58/proposed/#accessibility-requirements-mapping

Wilco: I was asked to look at how to present secondary requirements in our rules. It was added to the list of accessibility requirements with exactly the same language as before, which didn't work.
... What I came up with and looking for feedback is this format in the Accessbility Mapping section a Secondary Requirements section.
... See link posted for the added content.
... Removed the explanation of the secondary requirement from the background section.

Kathy: I like how you formatted the secondary requirements. The text might need some work since it's not always that the requirement is stricter than the rule. I can offer some suggestions on wording.

Trevor: I like it as well. Pretty clean and clear.

Lori: Grammar error. Replace this for these requirements.

Tom: Is that language of the bullet going to be freeform or a token?

Wilco: For now, I think it may be free-form.
... We'll explore when we talk about secondary requirements in awhile.

<Wilco> https://github.com/w3c/wcag-act-rules/pull/200

Wilco: I also added useful links section back in.
... Links allow you to see issues for the rule, open a new issue, or propose a change. It will only be on proposed rules.
... Also have a redirect so if you go to the URL without 'proposed' and the published one doesn't exist, will redirect to the proposed version.
... Published rules always have a proposed version, so the other direction isn't required.

Kathy: Those links are only on the proposed rules?

Wilco: Yes - not seeking changes on rules that are already resolved.

Better define how rules related to page states (30 min)

Trevor: Last time we talked about changing the approach to stateful rules to using the where...after language.
... Preface the applicability with this language. Felt cleaner.
... Started working on testcases that could go into a rule.
... Started some examples on how people might implement these.

<trevor> https://github.com/act-rules/act-rules.github.io/issues/1953

Trevor: I have two examples, three cases on how this works.
... First case, just enter text in input and show a status message. Second simulates a browser submission. Third is like a rolling log.
... Basic examples, but showing basic updates on a page so we can think about how rules might be implemented for this.

Kathy: I'm reading the applicability. "after activation of an HTML element". Could be activated not by a user?

Trevor: One of the bigger questions we have. Wilco mentioned they have a tool that watches interaction of a page. We were wondering if we should handle that, tool watches interaction. I set a timer for after this much time do a thing, and that's where that was coming from.

Wilco: I like these examples. I'm not sure I put that auto thing on a timer. I'm not sure I want to see both of these in a rule, but we can consider if we have some sort of listener that triggers. Sorry, a standard custom event that you could fire that an implementer could trigger to have the interactions happen.

Trevor: I thought about that. I wasn't sure if that was reasonable for implementers of the rule. I wasn't sure if we used these timed steps, like if the page loads, can't use straight timer.

Wilco: I'm curious because I'm not sure of other tools that watch a page and decide based on interactions that have happened. I'm curious is we're building a solution for one particular implementation.

Trevor: That I don't know.
... I've heard of some tools that will listen for people clicking through a webpage. Those tools may not be aware of ACT. But I've heard of other tools that use user's data browsing sessions to look for issues.

Wilco: Did a11y talks recently interview someone?

<Helen> https://a11ytalks.com/

Chris: On the event listener watcher type of thing to clarify. You mean for the automated piece, from a watching perspective, it's an automated capture, but manually watching? The output from a rules standpoint?
... We do something like that.

Wilco: So there is a use case for this.
... I quite like the idea of having a custom event listener that we are using for every test case that we use so that an implementer can do the interactions themselves or they can trigger the interaction
... We can explore that a little bit more. If we like that direction, maybe the next step is see if we can pull together a rule if that makes sense.

Trevor: I think I have most of the pieces here. Adding the testcases here becomes fairly trivial.

Wilco: Another thing we need to do is go through the rules format and figure out where if anywhere we need to make changes to allow this approach.

Trevor: I think personally I don't know anywhere in the applicability we really need to change is the exceptions.

Wilco: I would certainly would want to see an example added to the applicability. Not sure if there are any normative changes.
... What do you want to try next? Rules format update, or put into a rule.

Trevor: I think low hanging fruit is to put it into a rule.

Wilco: If have it by next Thursday, can share with the community group.

Link in context is descriptive

Wilco: This is related to the survey two weeks ago
... Q4, Trevor, one consistent implementation, but it's automated, so I question how it's measuring
... I don't know.

Trevor: I'm not saying it's impossible, but we've seen similar rules that automated tools have struggled with.

Tom: Wondering if they're using some heuristic that's not really testing the rule.

Wilco: Doing some clever stuff looking at Fail 5. I can ask SortSite.
... Determine if it's a valid implementation.

Lori: What was the company?

Wilco: Tool is SortSite, company is PowerMapper.

Kathy: Would it be helpful to have different types of failed examples that are not just one word?

Wilco: Maybe their heuristic is just failing things with only one word?
... Pass examples only have one word, so, not sure. I'll reach out to PowerMapper to confirm that this is consistent.
... I can ask how it's implemented.
... I'm sure there are some machine learning techniques that you can get a decent guess, but worth confirming.

Kathy: I don't want to bias against automated tools, but probably some examples that are more complicated than what we have here that might trip automated tools.

Wilco: I don't like the idea of tweaking the rule to make it difficult to implement simply because we feel it should be difficult to implement.
... If there are common types of failures that the rule doesn't describe or capture well, we should add. Not sure we should add simply because we think it should be hard for an automated tool to do.

<ChrisLoiselle> This is one rule, https://www.powermapper.com/products/sortsite/rules/acchtmllinktextreusedcontext/ not sure if there others they are referencing?

Trevor: I just want to make sure they haven't overtrained or oversimplified based on the testcases of the rule.

Wilco: We don't want tools to adjust their logic just to pass testcases, but we do want the to account for the scenarios of the testcases.
... This is a broader conversation. I'll talk to them and bring it back.
... Next comment, Catherine, NVDA reads Pass 3 period as "dot"

Tom: I think this is an AT issue. Looks appropriate.

Wilco: Might want to report this to NVDA, but don't think we should change the example.
... Next item, Catherine, N/A 1 and 3, extra a?
... Oh, that's the "a" element.

Tom: The color is a bit difficult to identify as a code element.

Wilco: We had decorated these in the past. I'm a little hesitant to change because all of W3C does it this way.

Helen: Not sure we should do that just because W3C does it.

Lori: Can we add quotes to the a so we know what the target of the sentence is.

Wilco: We can, but becomes inconsistent with how we do other things.
... We'd have to do it for all elements.

Lori: On my screen, there's not difference in the color, so it looked like a grammatical error.

Tom: Can we flag this as a separate issue and not block this rule?

Wilco: I think there's a design problem with the website. I will file an issue with the WAI website.

Lori: Seems like an accessibility violation

Wilco: I'll open an issue.

Kathy: While you're at it. I liked how the community group showed the code with a dark background. This is less appealing.

Wilco: I agree the styling is not clear. I'll raise it.

Helen: Before we move on, I can't find the minute notes from last week.
... I was going to create a PR in the about page section, was trying to find the notes.

Wilco: Q5, 2.4.9 should be secondary. Yes, should be fixed from earlier discussed change.
... Also comment by Wilco of an open issue.
... 1766
... Rule does not consider the title attribute.
... Q7, bracketed text in example descriptions.
... text in parentheses is difficult to read and seems not necessary.
... If I'm the only one, we can leave it.

<ChrisLoiselle> Regarding sortsite , https://www.powermapper.com/products/sortsite/rules/acchtmllinktextambiguous/ is on https://www.powermapper.com/products/sortsite/rules/accwcag2.1/ Wilco

Wilco: Looks like we need to sort out secondary requirements. Need to talk to SortSite. Resolve open issue.
... There are surveys open in two weeks.

Tom: I will be out in two weeks by the way.

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.200 (CVS log)
$Date: 2023/04/06 14:02:08 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision VERSION of 2020-12-31
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

Default Present: Helen, ChrisLoiselle, Wilco, trevor, thbrunet, kathy, present, catherine_droege
Present: Helen, ChrisLoiselle, Wilco, trevor, thbrunet, kathy, present, catherine_droege
No ScribeNick specified.  Guessing ScribeNick: thbrunet
Inferring Scribes: thbrunet

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


WARNING: No date found!  Assuming today.  (Hint: Specify
the W3C IRC log URL, and the date will be determined from that.)
Or specify the date like this:
<dbooth> Date: 12 Sep 2002

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]