W3C

– DRAFT –
ARIA WG

04 September 2025

Attendees

Present
aardrian, Adam_Page, Daniel, filippo-zorzi, Francis_Storr, front-endian-jane, giacomo-petri, Jacques, Jem, katez, keithamus, kiethamus, Matt_King, pkra, Rahim, sarah, scott, smockle, Stefan
Regrets
-
Chair
-
Scribe
aardrian

Meeting minutes

scribe aardrian

New PR Triage

jamesn: Skipping the first as editorial

jamesn: Skipping the second as editorial

jamesn: can @pkra look at this and push it if ready? w3c/aria#2619

jamesn: sounds editorial w3c/aria#2618

rahim: it is

jamesn: Can you mark it as editorial?

pkra: You can ignore my suggestion. We can do it some other time as a new issue.

rahim: Is the issue the single string, literal style? Want to be consistent with specs.

jamesn: I think `code` is what we normally use.

scott: This was introduced in Web Apps WG as recommended at the time.

scott: Needed `code` when formatting wouldn't work otherwise. So the specs are riddled with backticks as a result.

jcraig: It's not rendered that way. Since it's a new ReSpec thing, keep the backticks and replaces as we go.

jamesn: Is everyone ok with this? Brian?

bryang: Yup.

jamesn: We can make an editorial task for later.

jamesn: next is w3c/aria#2617

scott: The gist is the interest in an attribute.

scott: based on work from about a year ago to get this through OpenUI.

scott: Aaron is doing his best to copy the explainer.

jamesn: Do we need to agenda it?

scott: I'm all for talking about it, but timing.

jamesn: We'll take it offline. Anyone else want to review?

keithamus: I'll do it.

front-endian-jane: Me too.

jamesn: You should learn autocomplete, Adrian.

james: Last one is w3c/aria#2606

Adam_Page: It has two reviews. It's in accName.

james: Is it ready to merge?

Adam_Page: It has tests, so ready to go.

jamesn: Bryan or Melanie, can you merge this.

zakim nex item

zakim nexy item

New Issue Triage

<jamesn> w3c/html-aam#591

jamesn: w3c/aria#2622 is editorial as is w3c/aria#2621

jamesn: Rahim says JAWS is only one to announce title on `abbr`: w3c/html-aam#591

Rahim: SHould title contribute to accName? Can be verbose, but is not common experience.

Rahim: Should title map to accDesc?

<Jem> we have a ARIA-APG github issue for abbr, w3c/aria-practices#3319

jcraig: Every time there is a title, the browser should speak the title and that would be tedious.

jcraig: I pushed back on that.

jamesn: doesn't accName say it should use child content?

scott: This is under text-level element accName computation. Probably just needs to be clarified.

scott: Falls under title fallback.

scott: Touches on what I've brought up before for when does accName mean accName versus content?

scott: It's just the content. We've confused this.

<pkra> => w3c/aria#2569

scott: The `title` attribute was called out in a comment. JAWS has a setting to allow this, but probably doesn't belong in accName at all.

scott: should be probably mapped to something else.

scott: this is unique to JAWS

jcraig: VO does. It's mapped to help.

scott: Basing this off my last testing pass.

aardrian: I've tracked support, and it's variable. Not accName IMO

jamesn: Do we need more discussion or a PR?

Rahim: I'll do it.

jcraig: It seems we've come to agreement it shouldn't be accName.

jcraig: But there are platform mappings that are needed for HTML AAM

jamesn: next up is w3c/aria#2613

scott: I came here specifically for this issue.

scott: Lots of back and forth. When we created sectionheader|footer, we wanted them to be similar to group.

scott: Did not want to introduce these new roles and be too verbose. Announcing themselves everywhere when they were otherwise silent.

scott: Introduced because of a use case of naming a header/footer, that's how they became landmarks.

scott: But then people named them in other contexts, naming generic elements.

scott: This was to allow authors to use these as HTML intended and provide a meaningful role.

scott: There seems to be confusion for why they're exposed with only an accName — not accurate.

scott: There was never an expectation authors should start naming them to expose them.

scott; But if authors _do_ name them, they'll be exposed.

aardrian: Papering over a gap authors perceived in spec?

scott: That and handling things that weren't defined outside of specific contexts.

scott: Similar to groups and fieldsets — announced but not named.

jamesn: What needs to happen to move this PR forward?

pkra: Isn't the question what to do with the issue?

jamesn: I jumped ahead to the related PR.

scott: I updated a PR to align and correct for original spec intent.

scott: Also a response to giacomo's issue about naming them to expose the role.

jem: Usually a header landmark is a random or section header.

jem: YOu can have a heading structure within it.

jem: Why should it be mapping to section header versus just using headings?

Matt_King: I think you're asking an authoring choice. Why would they do this versus that. Feels outside scope.

jem: Scott's intention was an alternate header, but authors can use headings.

jamesn: The answer is HTML allows it.

pkra: And authors slapped names on them.

pkra: I think scott answered giacomo's question, so let giacomo answer it.

jamesn: With scott having to leave I wanted to jump to the PR.

giacomo-petri: I understand the intent. I'm fine with it. I'd expect an author using footer would use a section.

giacomo-petri: I would expect it to be communicated as a footer section.

giacomo-petri: Still not sure this should be exposed by AT as a grouping element.

jem: I like the comment that a header/footer in a section means something. I find the overall thing confusing, but understand authors will use it.

jamesn: Authors are using this where we wouldn't expect them, so exposing them everywhere would be a poor experience.

Matt_King: Exactly. Authors use lots of markups with no consideration for accessibility, though it has a side effect of verbosity.

Matt_King: We should assume an accessibility intent.

Matt_King: They might think adding the name could have some intent.

jamesn: Moving on...

pkra: Skip w3c/aria#2612

jamesn: Next is w3c/aria#2611

pkra: Came out of discussions and WebKit / APple statements.

pkra: w3c/aria#2604 came out of discussion in prior meeting.

WPT Open PRs

jamesn: ANything we need to look at jcraig?

jcraig: I dont' think so, haven't reviewed.

rahim: I think the array initialization is ready to merge.

Adam_Page: I'll review.

Adam_Page: James, should that pseudo content marker test be tentative?

jcraig: Good question. Don't know. Fine as tentative.

jcraig: If the results are fine,e veryone's mapping, and it aligns with the spect...

Adam_Page: ONly Firefox right now

jcraig: Then tentative and check bug trackers.

TPAC Agenda

jamesn: Need agenda items for TPAC. This is the old list of F2F candidates.

jamesn: Some may not be relevant.

jamesn: Start adding stuff. Add F2F Candidate label.

jamesn: This is driven by the groups, not chairs.

jamesn: Without topics, it won't be effective.

jamesn: Yes, we can enjoy beef. But ideally do some work.

jem: There is a WhatsApp group if it's your first visit.

jamesn: Ask Jemma for help!

jcraig: There are overlapping groups. She meant the general accessibility group.

Editorial: Add <header> (sectionheader) and <footer> (sectionfooter) html-aam mappings

jamesn: We kinda talked about this already.

jamesn: Rahim owns this. What do we need to do to progress it?

Rahim: Request to Giacomo and you to move ahead.

jamesn: I'm fine with it.

giacomo-petri: I'm fine with it.

Rahim: I'll ask scott to look.

Rahim: Ready to land

Daniel: I think this PR has changed to graphics AAM?

Rahim: No, they're Prettier extra comments. Is there more to do?

Daniel: Take out the main graphics.

Rahim: I'll remove those.

pkra: You might have to rebase.

Clarify relationship between aria-hidden and aria-owns

james: This is ancient.

Adam_Page: It's an old one. Consensus 3 years ago was to enshrine what Chrome and Firefox were already doing.

Adam_Page: That is the spec change. It has remained unchanged in 3 years, until introducing accessibility child / ancestor language, but updated since then.

Adam_Page: I also wrote a companion WPT test. 2 browsers have bugs.

Adam_Page: Not sure root cause nor how they relate to core of the change.

jamesn: Do we need bugs against browsers?

Adam_Page: Eventually.

Adam_Page: When WPT is published.

jamesn: Is that the next step to get this moving?

Adam_Page: Scott found a test case after `aria-owns` resolved, there was a heading element getting its name from its content before being displaced by `aria-owns`.

Adam_Page: So there was a named element devoid of contents. Not sure there is a capability to write any tests in WPT that can confirm if element has been moved in the tree.

Adam_Page: Can only test computed name.

jcraig: Correct.

Adam_Page: That doesn't need to be a blocker. Interesting quirk, tho.

Adam_Page: Added additional test case from that feedback.

jamesn: jcraig could you review this?

jamesn: Before we merge it, we need browser implementations.

Adam_Page: They're there, but with bugs.

Adam_Page: This not a super-essential feature. `aria-owns` and `aria-hidden` are a rare mix.

<smockle> w3c/aria#2577

clay: What's up with `ariaNOtify`?

ariaNotify

Jacques: The thing previously blocking us from shipping in Edge was positive feedback from this group.

Jacques: We got that, so we're waiting for another browser vendor.

jcraig: The position from other vendors got filled out. It's down to the nits now.

Jacques: The V2 has open questions that need more clarity. ANd it needs SR support, so not shipping soon.

jamesn: Apple can't tell us anything for timelines.

jcraig: Not actively working on it.

keithamus: WOuld love to work on it, can't promise timelines.

jamesn: THerefore not actively worked on by Gecko.

jcraig: This is whittled down to what can be implemented today by existing features.

jcraig: The downstream AT and API changes are blocking V2 features.

jamesn: But we would love to get this into the spec.

Jacques: Am I right in thinking it's getting those commitments?

jcraig: SOme tags we can add for awaiting implementation.

jamesn: The spec can note it's only in some browsers, but this is a bigger feature with positive statements, so the spec can move forward.

jamesn: That's a chair discussion we need to have. Ways of stating this editorially.

jamesn: Maybe a TPAC discussion.

Daniel: We would put this in the spec at the moment with Chromium and another engine. We could go to CR.

jamesn: Unless we're in permanent CR.

Daniel: Yeah, but we can take a snapshot.

jamesn: We need to persuade someone else to implement.

keithamus: If someone wants to have a go at adding it to FIrefox, I'm happy to pair with them.

Minutes manually created (not a transcript), formatted by scribe.perl version 244 (Thu Feb 27 01:23:09 2025 UTC).

Diagnostics

Succeeded: s/topic/Topic/

Maybe present: bryang, clay, james, jamesn, jcraig

All speakers: aardrian, Adam_Page, bryang, clay, Daniel, front-endian-jane, giacomo-petri, Jacques, james, jamesn, jcraig, jem, keithamus, Matt_King, pkra, rahim, scott

Active on IRC: aardrian, Adam_Page, Daniel, filippo-zorzi, Francis_Storr, front-endian-jane, giacomo-petri, Jacques, jamesn, Jem, Jem7, katez, Matt_King, pkra, Rahim, sarah, scott, smockle, Stefan