W3C

– DRAFT –
ARIA WG

26 June 2025

Attendees

Present
aardrian, CurtBellew, filippo-zorzi, Francis_Storr, giacomo-petri, hdv, jcraig, katez, np-at, pkra, sarah, Siri
Regrets
rahim
Chair
-
Scribe
hdv

Meeting minutes

New issue triage l-aam+repo:w3c/dpub-aam+repo:w3c/svg-aam+repo:w3c/dpub-aria+repo:w3c/graphics-aria+repo:w3c/graphics-aam+repo:w3c/mathml-

Daniel: we said we won't do triage every week, but every few weeks

Daniel: first one is from Peter, on title element mapping

pkra: I wondered if it is confusing that it isn't mapped… but it is also the name of the document in the spec

jcraig: looks like the link is different

Daniel: ah it's broken in the chat, you can follow it from the W3C agenda

pkra: if it is just closed because editors say there's nothing useful here, that's fine for me too

Daniel: let's wait for an editor's response

Daniel: then #2565, related concepts missing in aria-colspan and aria-rolespan

pkra: I took the liberty to tag as editorial and good first issue

Daniel: ok

Daniel: then #2556, from James Teh, should we still require ID for activedescendant to reveice focus

Daniel: there are some comments on this

pkra: the whole section is a bit dated… doesn't mention reflection at all… needs to be updated to pull that new, modern featureset in

pkra: it hasn't been touch in… forever

pkra: should update it to how it works in browsers. Anyone feel free to pick it up, also happy to do it

jcraig: I'd be happy to review

pkra: can assign it to me, and otherwise I'll bring it back

pkra: annevk would also be great to ask for clarifications re this

Daniel: then 2554, 2552, seem for editor?

pkra: yes mostly respec stuff

Daniel: then #2548 was from last week, we can skip for now and get back

PR triage 3Aw3c%2Fhtml-aam+repo%3Aw3c%2Fdpub-aam+repo%3Aw3c%2Fsvg-aam+repo%3Aw3c%2Fdpub-aria+repo%3Aw3c%2Fgraphics-

Daniel: there's 3 PRs, most editorial

Daniel: #2566 is mine, we can skip it

Daniel #2563, I think is being worked on?

pkra: I replaced two PRs that had been stale and needed updates, they were stuck… I rebuilt them and tried to incorporate the comments on the originating PRs and tagged the folks that already had reviewed

pkra: I hope this makes them easier to re-review. If I tagged you on one of those, please take another look

pkra: the fail is a false positive… am creating a link in a section that is RFC but is not actually changing it

pkra: re #2562… I think I was asked to help on this during editor's meeting… couldn't help because it is on Rahim's fork so made a new one

WPT PR triage

Daniel: James can you guide us through is there anything we should talk about?

jcraig: #52159 needs a review, is there someone here who would do a review? would be helpful

jcraig: it is literally just some file renaming

jcraig: the others are waiting for updates

Deep dive Planning

<Daniel> w3c/html-aam#584

Daniel: this one is scheduled for July 17th?

Daniel: yes

Daniel: remember, if anyone who wants to flag something for a deep dive, feel free to do so, you'd need to an intro

aardrian: seeing that the issue is closed and in it, and keithamus says in it his questions were answered, not sure if this still needs to happen?

pkra: we prepped it a bit last week

keithamus on chat: I think it still needs to occur

keihatmus on chat: i'll update with a comment

Are ARIA roles case sensitive or case insensitive?

Daniel: this is about case sensitivity of ARIA roles, we can discuss it a bit today

pkra: I think it's in the spec if you follow the thread in the issue. But it's not obvious, a clarification would be good

pkra: it's a little tricky because it comes down to HTML's implementation of ARIA

pkra: and I suppose SVG's

Daniel: should we work out what the clarification should be today?

pkra: Rahim opened an issue on WPT

pkra: if somebody wants to find a good way … I'm not sure what is a good place to do it, don't know if we want to prescribe to host languages how to treat their attributes

pkra: the WPT point is probably the best starting point

jcraig: a @@@ test should be able to verify that Gecko doesn't crash. Or a WPT test as long as that doesn't require the bit we haven't defined yet

jcraig: will add that as a comment

Daniel: ok, then we have a path forward

Presentational Roles Conflict Resolution (#2237)

Daniel: it seems to need a rebase

pkra: it has to do with Prettier, could be easier to start of fresh

Daniel: could you look at that giacomo-petri ?

giacomo-petri: yes

Address title attribute and name prohibited elemeents (#2292)

Daniel: it's already under review by Keith, Aaron, James and Rahim… Keith would you be able to have a look?

keithamus: yes

[html-aam] Addition: selectedcontent element (#2344)

Daniel: this is about selectedcontent, has a few reviewers

scott: probably needs input from Mozilla and Webkit folks

keithamus: I'll be the one implementing for Mozilla, you can add me

Daniel: ok

Clarify / expand value attribute mappings

jcraig: re previous topic: it still waiting for an official standards position from webkit

Update: legend to provide name to optgroup

scott: this is also about customisable select… would like people who will implement it to look at it

jcraig: haven't looked the actual feature yet

scott: all contingent on browsers actually shipping the feature

scott: so answer to Valerie on the GH issue is yes

keithamus: we have one implementation, which is Chrome. I've approved this PR so that could count as endorsement from Mozilla

scott: I'm personally inclined to merge all customisable select things at the same time, as they won't work separately

scott: need to figure out if anything has changed on these, haven't done anything on these for a while because of other priorities

jcraig: I added a link to the webkit standards position prerequisite… if you want to add a related issue on Gecko, please do

jcraig: there's a feature specific one and an accessibility specific one

scott: Option element mapping table (#2365) is also related

add aria-checked support to gridcell and row (#2390)

Daniel: is from sarah if I'm not mistaken

change default role for custom elements (#2383)

keithamus: I could review this, it looks like there are some reviewers already but not sure if people have looked at it yet

Fix incorrect role specified for html element (#2504)

Daniel: looks like it has 1 requested change from jcraig

jcraig: just resolved the thread on the PR

ARIA IDL updates (#2484)

Daniel: is there anything we can do now with this?

keithamus: changes are fine… but my requested change is that this is normative… and we need to review all implementors agree with the change, there are potential web compat issues so need an assessment from the browsers

keithamus: returning the reflected value isn't necessarily true today… so while that's fine… Salesforce has a polyfill for aria mixin, when I talked to them last about this, they said they rely on the existing 'broken' reflecting behaviour

keithamus: may want to investigate it further to make sure there are no issues

jcraig: would you suggest we write a tentative WPT test based on this change?

keithamus: would be an excellent next step

clarify parent-child relationships for menu, group and menuitem (#2483)

pkra: I put this on the agenda recently so we can talk about it… the donut scoping thing… see if we can find some good language to describe that

Daniel: ok so will be on an agenda of a meeting in the future

pkra: that's my hope

Empty or non-matching values (#2235)

Daniel: I see aardrian and filippo-zorzi on it,do you think you can look at it?

aardrian: can do today

filippo-zorzi: me too

clarify img mappings (#2221)

scott: there was an issue re that img element could have either src or srcset, you don't necessarily need either. This PR was done to match HTML

scott: I hope this can go in now, not sure what else needs to happen at this point

pkra: just reflects reality, right?

scott: yes, that's also what the WPT tries to show, I put together a variety of img use cases and the various ways to give accessible names

scott: not sure what else is needed than approving reviewers

pkra: jcraig, you commented on the original PR… the intermediate comments have been resolved, only some comments on the WPT

jcraig: will review

Add normative text on calculating aria-level, aria-posinset, and aria-setsize (#2104)

Daniel: still has conflicts to resolve

sarah: I can resolve the conflicts, and will have a look at the comments

pkra: there's also the bigger question that jcraig raised re headings without a level

pkra: but this part doesn't talk about heading level

sarah: this was originally written about trees, could be written explicitly

jcraig: don't necessarily think we should move forward with the default value change, unless heading is explicitly opted out of this, in the same PR. Because it will result in more user tedium, odd user tedium too

jcraig: in the grand scheme of things it's not the most important accessibility nag I have, but I do foresee user impact and increase of tedium for the user

pkra: when I tested, NVDA and JAWS were exposing @@@

jcraig: in VoiceOver we try not to add more verbosity when not necessary, eg some Windows screenreaders say Tab when you press Tab while the user probably already knows they pressed that

sarah: agree with you… but this PR has tree groups, we could make it more explicitly

jcraig: does the user presented language affect the way tests are written? if the tests are written with headings undefined, that's how we need it written

sarah: I think it's already explicitly not headings, but will double check the wording to make sure

jcraig: one of the reason IDL project has been so complicated is prose based exceptions that are all throughout ARIA since many years

sarah: it's prose but written as an explicit set of logical steps

sarah: this was initiated by my desire to make flat trees possible… but we got somehwat sidetracked

keithamus: fwiw for the IDL stuff; if the attribute is not present the IDL should reflect now, don't think it would matter with regards to IDL

jcraig: so this particular change would not affect IDL at all, just host language IDL implementation?

keithamus: yes

jcraig: that's manageable then

sarah: @@@, the document outline was never really a thing

jcraig: based on keith's comment I think my concern is less relevant

Daniel: so you'll look at the wording Sarah?

sarah: yes

jcraig: we can also add rahim

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

Diagnostics

Succeeded: s/peter/pkra

Succeeded: s/irc/chat

Succeeded: s/mathcing/matching

Succeeded: s/tehn/then

Maybe present: Daniel, keithamus, scott

All speakers: aardrian, Daniel, filippo-zorzi, giacomo-petri, jcraig, keithamus, pkra, sarah, scott

Active on IRC: aardrian, CurtBellew, Daniel, filippo-zorzi, Francis_Storr, giacomo-petri, hdv, jcraig, katez, np-at, pkra, sarah, Siri