W3C

Accessibility Conformance Testing Teleconference

24 Mar 2022

Attendees

Present
Wilco, Will_C, thbrunet, kathy, trevor, Helen, Todd, Daniel
Regrets
Chair
Wilco
Scribe
Will_C

Contents


<Wilco> scribe: Will_C

Will: acknowledges he is the scribe

Update WAI ACT website

<Wilco> https://www.w3.org/WAI/standards-guidelines/act/rules/

Wilco: this page is different now. Banner has been updated.
... Sidebar has been updated with new look to page content

Trevor: Are we going to work with the deprecated section? It's getting big. Wilco: yes

Wilco: How do we feel about the navigation at the top now?

Kathy: It's a bit odd in the banner

Wilco: It went live without our review. it's got a UI bug even.

Helen: the formatting is interesting. It's not conventional.

Will: said stuff.

Wilco: Updated all proposed rules with links to test cases
... links within test cases are broken, b/c absolute paths, we are working on it.
... automatic updating is occurring for the W3c website. goes live every couple of days
... next, working on automating the implementation. Trialing it on axe-core.

Helen: work faster, wilco
... just kidding

Wilco: now publishing proposed rules alongside approved rules

ACT rules sheet and Survey Results

Wilco: akn7bn

Helen: Status update - looked at the PR and fixed the items in it. out for review. might not have done everything.
... Don't know if karen did more examples

Open ACT pull requests

Wilco: 1817, out for reviews
... 1808 into call for review
... 1805 - jean yves asked for reviews
... proposing a new section at the top of test cases where the content of test assets can be seen. Has mocked up a look
... Not merging it, just approving concepts
... 1804 - trevor to look ar that
... 1794 - needs reviewers, will and kathy added

WIll C: too much coffee

Wilco: update on widget role.
... conclusion is that we cannot do this, due to new definition of 'consistent implementation'
... you can't fail something for 4.1.2 and then also fail it for 2.4.4. for example
... does everyone feel like revisiting?

Helen: 10 years down the road, it'll have to be revisited. Easier to keep it per item, images for 1.1.1, links 2.4.4., buttons 4.1.2.

Wilco: assuming it's cool
... Helen, 1779 is approved
... Rest of pull reqs are old. skipping those
... gonna talk to the community group chairs about closing old PRs. At the least, non-members

Survey response: Element with aria-hidden has no focusable content

Wilco: act rules check - didn't complete. Passed example 4 needs a clearer example
... 1807 is an open issues. Failing items with negative tabindex and aria-hidden=true
... Where does everybody stand? I don't think things with a negative tabindex should always fail when there is an aria-hidden="true"
... it's not alway a failure

kathy: Question of wether you could still use your mouse to activate the button?

Wilco: yes if you're not obscuring it

helen: you don't know in the context, sometimes things get hidden visually for a more accessible route

Will: talked a lot and it didn't matter in the end.

Kathy: If we can update the TC so that it isn't clickable then it wouldn't be a question

Wilco: Then we have to update the rule
... there is an assumption in all rules about alternative versions to be ok
... Tabindex-1 makes something focusable, making this tricky.

Tom: People typically do it to take it out of sequential order.
... The more important rule is don't have tabbable elements inside an aria-hidden

Wilco: Proposal - we add an assumption saying that focusable elements that are not part of sequential focus navigation can be ignored, acknowledging that there might be other acc issues
... might have to adjust the rule to focusable elements that are part of sequential navigation

Tom: Change the name from focusable to tabbable?

<dmontalvo> +1 to those

Wilco: Any opposition to that?

Tom: THe argument is if you can see it to click on it you can see it to read it

Daniel: Question on definition of focusable. Tabindex-1 makes it focusable but it's not focused. Can we make it clear we mean 'programmatically focusable"/

<Wilco> draft RESOLUTION: Accept the case in 1807 as inapplicable for the aria-hidden rule and add an assumption

Wilco: no opposition

Helen: show me where skip links hurt you

Wilco: No

no objections

RESOLUTION: Accept the case in 1807 as inapplicable for the aria-hidden rule and add an assumption

Wilco: updating the spreadsheet now.
... tom's comment about pass 4 making it difficult to automate
... automatically redirecting focus does make it more difficult to automate
... separate issue open for that.

Helen: Wan't Jean-yves talking about this one?

Wilco: 1811 covers this very thing.

Tom: i struggle because if it isn't exactly done this way, I have no idea how you put focus on it, script-wise

Wilco: Selenium lists out event listeners, but the DOM doesn't
... difficult to test, not a realistic example.
... created as a way to keep focus in the modal.
... would prefer that modal case as example over this simplified version

Tom: Every one of the failed examples are eventually a 'cant tell'

Wilco: how do others feel?

Helen: Whatever the devs say, I back up.

Will: ditto

Trevor: We have rules that are difficult for automation, manual testing will catch, but agree this one is not realistic

Wilco: don't throw out, adjust

Tom: illustrate it better. show an event listener hook

Wilco: How about replace it with a modal pass and modal fail that show redirecting and focus in the background?
... this came from a real-world false positive

Daniel: Modal will clarify this

Tom: Is there another role that could be applied other than link?

Wilco: proposal - replace passed ex 4 with a pass and fail example of a modal

Agree

No objections

<dmontalvo> +1

RESOLUTION: Replace passed example 4 with pass and fail example of a modal

Wilco: going through other issues
... conclusion it needs more work, updating tracker
... sorry Kathy

Kathy: might be too technical with the modals

Tom: I can at least come up with the modal example

Survey response: audio element content has text alternative

Wilco: audio alternative - from tom: should there be assumptions regarding usable native controls.

e7aa44

Tom: every example just uses 'controls' attribute

Wilco: I can see us adding a test case that uses a button element

Will: has value from a testers perspective since we want to see issues that match what we have found

Wilco: Blocker or no?

no blocker

Wilco: from Jennifer- the incorrect text took a minute to spot

Kathy: I didn't have a problem spotting the problem, but was amused

Wilco: does this need an adjustment?
... we can adjust, not a blocker?

Not a blocker

Wilco: 1346 - seems like a thing we may wanna do
... definitions don't belong in a note. play button should be a link to the glossary

WIll: +1

Daniel: +1

Wilco: ended on q6 - #1346 needs to be resolved
... failed examples to be more obvious
... add example of non-native controls
... last two are optional
... there are two new surveys open: meta-element and meta-viewport. Please fill out before next week4

END

Summary of Action Items

Summary of Resolutions

  1. Accept the case in 1807 as inapplicable for the aria-hidden rule and add an assumption
  2. Replace passed example 4 with pass and fail example of a modal
[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.200 (CVS log)
$Date: 2022/03/25 10:58:51 $