W3C

– DRAFT –
ARIA WG

09 January 2025

Attendees

Present
aardrian, aaronlev, Adam_Page, bryan, Daniel, filippo-zorzi, giacomo-petri, jamesn, jcraig, katez, keithamus, lola, pkra, Rahim, ray-schwartz, sarah, smockle, spectranaut_, StefanS
Regrets
-
Chair
spectranaut_
Scribe
jcraig, smockle

Meeting minutes

New Issue Triage

w3c/aria#2409

spectranaut_: shadow dom... left a comment for response... may close

pkra: close it

w3c/aria#2406

spectranaut_: Children presentational true element nesting another control injected via JavaScript

giacomo-petri: there's a linked issue, possibly a dupe... will check

w3c/aria#2404

[ARIA] Requirement to ignore landmark roles without names creates infinite loop

should clarify... scott?

scott: we haven't identified an algorithm as a WG. Implementors have and may share.... should clarify for mike.

aaronlev: will share blink's strategy

scott: not as much about the name; about exposing a role basewd on the existence of a name.

scott: will tackle them both... need implementor to help.

w3c/core-aam#244

Update AX API mapping for aria-multiselect

spectranaut_: I will update

w3c/aria#2403

Clarify use case for aria-selected for 'grid' role descendants

ask reporter if this is addressed already

sarah: i don't think it would be

scott: i thought his issue may be resolved by related discussion in sarah's PR

assignee? sarah... (thanks sarah!)

w3c/aria#2402

Should the 'row' role really be necessary for parents of 'gridcell' and other cell role elements?

spectranaut_: scott replied... dupe?

reporter investigating based on scott's comment

New PR Triage

w3c/aria#2410

editorial: revise custom stylesheets

pkra: editorial. fyi for editors: removing core-aam.css from core and pdf

w3c/aria#2408

editorial: refactor aria.js

pkra: editorial. in draft

w3c/aria#2407

[aria] Drop “deprecated on this role” attributes from per-role tables

pkra: this request changes aria.js so we may need to wait. also not sure this is desirable. I left a comment.

w3c/aria#2405

Tentative: update point E. Host Language Label of the Acc Name algorithm

<giacomo-petri> w3c/accname#108 (comment)

closes an old issue I've linked above in IRC

issue causes ambiguity that i've tried to correct with this PR... (enumerates specific issues)

spectranaut_: reviewers james, bryan, and scott

any new reviewers

lola: i can

WPT Open PRs

Deep Dive planning

spectranaut_: whitespace deep dive needs to be rescheduled... afternoon so jamie teh can attend. i will reach out to mel for scheduling

Discussion of deprecation

<spectranaut_> https://github.com/w3c/aria/pull/1862#pullrequestreview-2028023050

jamesn: deprecation is not how ARIA has done things in the past.

jamesn: deprecation could result in implementation REMOVAL which has downsides

<aardrian> +1 to scott

scott: agree with jamesn, deprecation has its place, but this is not it.... do not need a long tale list of deprecated features

sarah: also agree

jamesn: i understand Wilco's concern for the validator, but they should be able to manage it

The content-visibility CSS property across browsers

giacomo-petri: content-visibility accessibility per role is managed differently across engines... especially content-visibility: auto;

tricky which to expose in which scenario

jcraig: deep dive topic?

aardrian: question... if it's hidden from a engine ax tree, does it also not participate in accname?

giacomo-petri: ambiguous what to do, but webkit behavior is different from gecko/chromium

spectranaut_: wpt test?

jcraig: You’re saying `content-visibility` (particularly `auto`) is ambiguous; are differences a bug (e.g. in Webkit)?

giacomo-petri: I’m not saying one implementation is wrong; only noting inconsistency.

giacomo-petri: Testing is challenging because `content-visibility` behavior is viewport-size-dependent.

jcraig: We don’t know the window size. So, instead, we could test a fixed-size (e.g. 100px) scrollable div—and we could know how many items to expect in it.

scott: Anecdotally, this has come up twice: Once with OpenUI re: an `aria-expanded` toggle button. Once at Microsoft, where there was confusion about whether hidden content is `display: none`.

Rahim: Can we add a CSS-AAM label to this issue?

spectranaut_: This issue should be moved out of the core-aam repo.

<Zakim> jcraig, you wanted to ask how it works with non-AT in-page search

<spectranaut_> new issue link: w3c/html-aam#572

jcraig: Does the spec say in-page search should find elements hidden with `content-visibility`?

jnurthen: Yes, it does.

jcraig: Because all of these, including `content-visibility: hidden`, show up to in-page search, they should also show up to AT.

<aardrian> +1

<giacomo-petri> content-visibility: hidden improves on both of these options. Because the contents are skipped, the user agent isn’t spending time on them when they’re not active. They’re also not visible to screen readers, find-in-page, and other tools. And because user agents should preserve previous styling/layout work if possible, if the view was

<giacomo-petri> displayed before, re-rendering it might be very fast.

scott: The `<details>` element in the collapsed state uses a version of `content-visibility: hidden`, so content is removed from the accessibility tree and visually hidden while remaining findable by in-page search. We don’t want to break find-in-page or start showing content in the accessibility tree that is supposed to be hidden.

<Zakim> giacomo-petri, you wanted to reference https://www.w3.org/TR/css-contain-2/#using-cv-hidden

giacomo-petri: Does the CSS Containment Module Level 2 (linked above) contradict what we’ve been saying?

jnurthen: Which value—`hidden` or `auto`?

giacomo-petri: `hidden`.

<Zakim> jcraig, you wanted to state maybe there is an ARIA-AT like expectation here too for <details>; if it should be hidden but found in search, it should be for aT search too

jcraig: If the expectation is that `<details>` is visually hidden but findable by browser in-page search, then it should also be findable by AT search. `content-visibility: auto` should be findable by (e.g.) skip-to-next-heading.

spectranaut_: This is a good place to stop this conversation, and we can pick it up again in a deep dive.

spectranaut_: We’ll have a deep dive on January 23rd.

[accname] Explicitly state UAs must ignore “aria-label” for slots

spectranaut_: reminder we need to review this.

Keith: I’ll review it

Minutes manually created (not a transcript), formatted by scribe.perl version 242 (Fri Dec 20 18:32:17 2024 UTC).

Diagnostics

Succeeded: s/identified an algorithm/identified an algorithm as a WG. Implementors have and may share./

Succeeded: s/i ca/i can/

Succeeded: s/keith: i will review. apologies.//

Succeeded: s/spectranaut_: gentle ping for keith/spectranaut_: reminder we need to review this./

Succeeded: s/.zakim next item//

Succeeded: s/zakkim, next item//

Maybe present: editorial, jnurthen, Keith, scott, Tentative

All speakers: aardrian, aaronlev, editorial, giacomo-petri, jamesn, jcraig, jnurthen, Keith, lola, pkra, Rahim, sarah, scott, spectranaut_, Tentative

Active on IRC: aardrian, Adam_Page, dmontalvo, filippo-zorzi, giacomo-petri, jcraig, katez, pkra, Rahim, sarah, scott, smockle, spectranaut_, StefanS