W3C

– DRAFT –
ARIA WG

16 April 2026

Attendees

Present
aardrian, Adam, CurtBellew, Daniel, filippo-zorzi, Francis_Storr, giacomo-petri, katez, melsumner, pkra, scott, Stefan
Regrets
-
Chair
ValerieYoung
Scribe
pkra

Meeting minutes

New Issue Triage

<github-bot> I can't comment on that because it doesn't look like a github issue to me.

spectranaut_: html-aam 604. li outside list

pkra: it would be great if implementors could implement this change

spectranaut_: right. looks like we have a bunch of these.

scotto: was wondering about this. Conversation with chrome engineer, that they do look at WPTs and track failing ones to deal with.
… but of course the WPT has not been merged either.
… still, it would be good if we only need one.
… WPT or browser bugs.

keithamus: just to say: there's a lot of WPTs and a lot of failure and thus a lot of noise for teams, at least for us.
… so it's worth filing a separate issue to make sure it's surfaced with the right team

scotto: thanks for clarifying!

spectranaut_: if you're interested, we could set up a deep dive for people who want to learn filing browser bugs.
… though I wonder why the WPT test isn't merged.

cyns: maybe we should have separate processes for different browsers.

keithamus: I suspect they're similar though.

dgrogan: we have the same firehose problem as firefox. If you can find the automated WPT bugs and ping it, that would be sufficient.

spectranaut_: I'll look into landing the WPT PR.

keithamus: I think WebKit doesn't have automatic bugs for WPT failures.

spectranaut_: aria 2767. listbox and accname

scotto: listboxes as popups often don't have accessible names. I haven't seen anyone being impacted by this. I don't think testing flags it since combobox names cover it. Conversations with other developers made me wonder if this requirement should be more nuanced.

giacomo-petri: to echo what scotto said. Our tool triggers an error for this kind of scenario and I agree with Scott that this change would negatively impact users.

scotto: I'm not sure how this works from the automatic table generation code in the spec. so maybe not a great first issue.

pkra: there's a wider issue on handling more nuanced requirements in the automation. I think it's a good first issue in that we editors should work out the automation.
… I'll tag the issue that reflects this.

giacomo-petri: just fyi, I think the suggestion role and separator are cases.

mattking: I'm wondering if this kind of situation should also look into accname, i.e., whether this should affect calculation.

melsumner: I can take it on.

bryan: me too.

spectranaut_: next up css-aam 17.

aardrian: I just wanted to flag it quickly so it's not lost.

spectranaut_: should we agenda it?

aardrian: I'm not sure anymore.
… but I assigned myself to update.

spectranaut_: great.
… next graphics-aria 14

pkra: I just wanted to put this down. inspired by recent svg-aam text discussions and longtime on my mind.

cyns: assign to me so we can track as well.

spectranaut_: next aria 2762

pkra: it's a bit of a mismatch. Definition doesn't suggest any specific way to connect to term. Term suggest something. We should clarify, probably a good first issue.

New PR Triage

<github-bot> I can't comment on that because it doesn't look like a github issue to me.

spectranaut_: just one from peter

pkra: still a draft. I'm working on it.

Deep Dive planning

<github-bot> I can't comment on that because it doesn't look like a github issue to me.

spectranaut_: nothing new.

WPT Open PRs

<github-bot> I can't comment on that because it doesn't look like a github issue to me.

ARIA in HTML for revised recommendation, please have your AC reps vote

spectranaut_: just a reminder to please ask your AC rep to vote.

CSS scroll-marker-group modes && WAI-ARIA patterns

<github-bot> I can't comment on that github issue because it's not in a repository I'm allowed to comment on, which are: w3c/aria w3c/aria-common w3c/core-aam w3c/accname w3c/html-aam.

daniil: to recap, CSS scroll-markers were introduced, in particular for native carousels.
… this is our update after feedback from the ARIA WG
… in my explainer I compared the two modes for scroll-marker-groups.
… the first one is tab mode. the second links mode.
… I would suggest to go through the explainer a bit.
… or wait until people have read it?

dgrogan: could you live share some demos perhaps?

daniil: yes. First one is tabs mode.
… one page of the carousel is one tab. Only the active tab is present in the acc tree.
… this was one change - browsers should now auto-hide it.
… here's a codepen demo
… I can tab to get to the first marker. then left/right arrows moves through the tabs.
… in the a11y tree you can see that only the active tab is in the tree.
… when we're at the marker, we go from the marker to content.
… but another tab won't take you to the next tab but out of the group.
… the second one is links mode
… in this mode, every scroll marker is a tab stop. They act like a focusgroup.
… another codepen demo
… we go to the first link. Tab moves me through the links.
… activation moves me to the content.
… just to recap, we discussed aria-current a while back and it's out of scope for this explainer.

spectranaut_: thank you.

mattking: I will need to read the explainer but it sounds like the first case is very similar to tabpanel behavior so I hope we can align there.

<Siri> +1 to Matt

daniil: we followed the tabbed carousel pattern.

mattking: great.

daniil: the links case is from the landmark navigation pattern.

cyns: there was a proposal a few years back from Google on a customized tab panel.
… is this a continuation of that?

daniil: I don't know but don't think so. This sounds like HTML but here we are talking about CSS

scotto: I think it's definitely different.
… it sounds like a lot of the feedback Sarah H, Sara S, myself and others provided is being considered.
… I would suggest to please update the codepen to include focusable content. Otherwise it's not as clear what the keyboard behavior would be.
… it will take me a while to take a look.

mattking: I'd second that. Also something before and after to make it easier for AT users.

daniil: thanks, will do.

<sakhapov> Rob's examples - https://flackr.github.io/web-demos/css-overflow/scroll-marker-type/

spectranaut_: it would be great if many people could take a look.

<cyns> I think this is the proposal I was talking about https://css.oddbird.net/toggles/explainer/

spectranaut_: do people know how to take a look? E.g. get chrome canary and activate experimental features.

<dgrogan> chrome:://flags

<dgrogan> oops. chrome://flags

<sakhapov> chrome://flags/#enable-experimental-web-platform-features

mattking: yes more details for testing would be great. Also in the issue itsef.

spectranaut_: right. that would be good.

siri: regarding link mode. tab mode makes sense. But also in the linked mode arrow keys is the expected way?

daniil: for links you just tab through them.

siri: but in the example I thought I had to

daniil: for me arrow navigation doesn't work.

siri: curious, it's different for me. I'll look into it.

Updates to aside element

<github-bot> I can't comment on that github issue because it's not in a repository I'm allowed to comment on, which are: w3c/aria w3c/aria-common w3c/core-aam w3c/accname w3c/html-aam.

spectranaut_: there was a recent comment from APG putting it on the agenda.

Daniel: from APG user research, chrome and firefox mark aside as complimentary in header and footer. But spec says only main and body, unless named.
… we should either match implementations or file bugs.

spectranaut_: this is testable already, right?

scotto: there are tests from 2023.

mattking: I don't think we need to re-discuss. Browsers are just not doing it.

scotto: looking through the tests from 2021. In 2023 James Craig commented on updated tests.
… there's a lingering PR of mine that just wasn't merged.
… due to the implementation mismatch.

spectranaut_: I'll look at WPT tests and follow up - unless someone else wants to.

mattking: sounds good. Daniel?

Daniel: yes. It sounds good to update tests to get things moving.

mattking: we have a PR in APG to say the same thing that html-aam says. But then we noticed the discrepancies and thus wanted to clarify.
… which brings us back to the start of the meeting.

spectranaut_: agreed.

Minutes manually created (not a transcript), formatted by scribe.perl version 248 (Mon Oct 27 20:04:16 2025 UTC).

Diagnostics

Maybe present: bryan, cyns, daniil, dgrogan, keithamus, mattking, scotto, siri, spectranaut_

All speakers: aardrian, bryan, cyns, Daniel, daniil, dgrogan, giacomo-petri, keithamus, mattking, melsumner, pkra, scotto, siri, spectranaut_

Active on IRC: aardrian, Adam, CurtBellew, cyns, Daniel, dgrogan, filippo-zorzi, Francis_Storr, giacomo-petri, katez, keithamus, melsumner, pkra, sakhapov, scott, Siri, spectranaut_, Stefan