W3C

– DRAFT –
ARIA WG

29 May 2025

Attendees

Present
Adam_Page, ChrisCuellar, filippo-zorzi, front-endian-jane, giacomo-petri, katez, keithamus, Rahim
Regrets
-
Chair
-
Scribe
jamesn

Meeting minutes

- WPT Open PRs

- Deep Dive planning

- No meeting June 5th

- Accessibility Compat Data Deep Dive Proposal

- Update: fix inconsistency around aria-setsize=-1

- PR review requests?

<spectranaut_> join the deep meeting link: https://www.w3.org/events/meetings/ef15df01-f812-4d5e-a91a-f2c6b0a5a7ed/

<spectranaut_> shoot that didn't work

<giacomo-petri> :D

<spectranaut_> I'll try something else

<keithamus> Can we use the prior deep dive meeting?

<spectranaut_> no its the same link

<spectranaut_> and it won't let us join that meeting

<spectranaut_> I'm going to add a new link to the current meeting

ha - the host has another meeting scheduled

<scott> sort of present ~

Daniel can only have 1 and the ARIA one is scheduled

<giacomo-petri> <scott> sort of present ~ :D:D:D

<spectranaut_> Got it, ok, I added a new link to the current aria meeting

<spectranaut_> https://www.w3.org/events/meetings/5a155237-d896-464b-9c5f-6dd1654293ae/20250529T130000/

<spectranaut_> load this page and click "join now" ^

New PR Triage New Issue Triage

<front-endian-jane> attending+

WPT Open PRs

Deep Dive planning

WPT Open PRs

jamesn: sounds like this new accname Issue is not urgent

spectranaut_: assigned to mel

No new WPT PRs

Deep Dive planning

spectranaut_: just had a deep dive on sort order. Daniel can provide recording if you want it

spectranaut_: Adrian will provide a write up

No meeting June 5th

Accessibility Compat Data Deep Dive Proposal

Update: fix inconsistency around aria-setsize=-1

spectranaut_: This is an old PR about aria-setsize

spectranaut_: Adam volunteered to research

Adam_Page: Did some research to see what browsers are doing

<spectranaut_> link to research: w3c/aria#2341 (comment)

Adam_Page: with this PR we are formalizing the language over what it is. Any non empty string that is not a positive integer. 0 & -1 should be indeterminate

Adam_Page: Says now user agents must not caluculate

Adam_Page: created tests - and then listened to browser / screen reader combos

Adam_Page: also for a number of different controls

Adam_Page: all of the browsers exhibit failures

Adam_Page: found bugs for FF, Rahim created one for webkit, created a new one for Chromium

Adam_Page: then added those links to the PR

spectranaut_: thanks for the research

Rahim: +1 thanks for the investigation

Rahim: prepared a patch. My question is about the indeterminate size of a set and what the AT experience should be

<scott> i appreciate adam just outlining to everyone what he did, as i'm sure the steps to test/file bugs may not be known to everyone

Rahim: what is the expected AT experience

Adam_Page: purposely didn't give any expectations

scott: started a few years back for tabels and grids which are loading "awesome" data that everyone cares about. Getting bugs that things were all over the map.

scott: zeroed in on the idea that if we don't know how much stuff is there - say it is indeterminate or dont' say a count that is not accurate

Adam_Page: there were a few cases where I marked a few use cases as satisfactory

Rahim: in terms of the scenarios where there needs to be a repair done. Could the author override the number of items in a set with set size. Could they put set size to -1

scott: that is a signal that they are trying to say that they expect more

if someone uses it incorrectly then that is incorrect

<spectranaut_> jamesn: I'd like if we could suggest screen reader information

<spectranaut_> jamesn: such as "at least 5 rows"

<spectranaut_> jamesn: where 5 is the number of rows currently

<spectranaut_> Rahim: would that be confusing if it changes?

<spectranaut_> jamesn: well we know it will be adding more

jane: don't disagree it may be usedful to say there are at least this many - if the number calculated is wrong

jamesn: yes - i'd suggest that it uses the max number of the aria-rowindex

<scott> scott says "argal bable fagabble and everyone says "ok" to that"

jane: believe you are seeing enough bugs in the wild that don't want to encourage screen readers to calculate any values

PR review requests?

(live PR updates)

<spectranaut_> https://deploy-preview-2105--wai-aria.netlify.app/#accessibilityroleandproperties-correspondence

<keithamus> w3c/aria#2541

<keithamus> https://deploy-preview-2541--wai-aria.netlify.app/#accessibilityroleandproperties-correspondence

<scott> posting this PR here as it has all accepting reviews, but one waiting from jamesn - w3c/aria#2416 :D

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

Diagnostics

Maybe present: jamesn, jane, scott, spectranaut_

All speakers: Adam_Page, jamesn, jane, Rahim, scott, spectranaut_

Active on IRC: Adam_Page, ChrisCuellar, filippo-zorzi, front-endian-jane, giacomo-petri, jamesn, katez, keithamus, Rahim, scott, spectranaut_