W3C

– DRAFT –
ARIA WG

12 December 2024

Attendees

Present
Adam_Page, alisonmaher, filippo-zorzi, Francis_Storr, giacomo-petri, hdv, katez, pkra, sarah, scott, smockle, spectranaut_
Regrets
-
Chair
-
Scribe
Adam_Page

Meeting minutes

<giacomo-petri> present

New Issue Triage

spectranaut_: aria/#2397
… dupe, I’ll close
aria#2396
… editorial, skip
svg-aam#38
… anyone in the WG interested in working on SVG?
… there are a lot of incorrect mappings, and we’re well-positioned to fix them

jamesn: we’ve taken ownership of svg-aam

<dmontalvo> New ARIA Charter

jamesn: would be ideal to get someone from the former WG to own this

<smockle> https://www.w3.org/2025/01/aria-charter.html#history

smockle: if this is something we’re owning, we should fix in our charter: “joint deliverable”

dmontalvo: I don’t think there’s an issue in the current charter, I’ll confirm

pkra: I have some interest in SVG mappings

jcraig: I filed some issues in 2023 re: WPT testing for graphics-aam and related svg-aam
… just got stalled because of access

spectranaut_: it’s in our repo now, so we should have access

pkra: how about a deep dive?

dmontalvo: it’s true that the current charter still says “joint deliverable”, but don’t think it will be an issue to just leave as is

New PR Triage

spectranaut_: aria#2395
… from scotto
… consistent naming for form controls

scotto: html-aam is not sufficiently descriptive when talking about gathering name from subtree
… and also there’s an accname issue about form controls nested within labels
… their values don’t get used in the name computation
… so I incorporated that into these rewritten steps
… noted this as editorial. It’s not changing any implementation, just improve accuracy of the wording

spectranaut_: any reviewers?

jcraig: I will

hdv: I will also

BryanGaraventa: I will also

spectranaut_: we should be able to land this soon
core-aam#241
… searchbox mpping for ATK/AT-SPI2 to match IA2
… need reviewers

filippo-zorzi: I will review

scotto: do we want to add a label to resurface open PRs?

spectranaut_: yes, add `agenda+` and a comment
… say “it’s just for review” and we’ll put it in the right agenda section

pkra: we also have the GitHub summary email — maybe we could tap into that

spectranaut_: ah yes, the chairs could review that and choose things to put in the agenda?
… we’ll think about that
… if you ever need movement on something, just add `agenda+` and a comment

WPT Open PRs

spectranaut_: wpt#49647

giacomo-petri: I’ll review

hdv: I’ll review too

Adam_Page: me too

jcraig: we only need 1 reviewer for WPT and then ping me

spectranaut_: wpt#49645

scotto: this one is trickier since it’s only for the customizable select
… can’t be fully tested in all browsers yet
… but hoping to get the ball rolling

spectranaut_: should this be linked to an issue/PR?

scotto: yep, it’s linked further below

jcraig: in GitHub, go to Checks > wpt.fyi > any of the browsers > Latest results to see a visual comparison
… I’ll check #49647 now
… it can land as tentative

spectranaut_: any other reviewers?

sarah: I’ll review

spectranaut_: wpt#49643

scotto: this is closing the loop on the linked PR
… we rolled back contextual roles for listitem
… but there was still aria#2212
… and browsers are all doing different things for one text
… so made this WPT test to record

Adam_Page: I’ll review

spectranaut_: wpt#49552
… we talked about this last week
… no other reviews needed

jcraig: I’ve got a minor comment for scotto

scotto: I’ll review

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

spectranaut_: new PR from a couple weeks ago
… keithamus volunteered to look into it
… let’s skip for now and talk with him at the next opportunity

scott: I agree

[AriaNotify] Naming for the priority property

spectranaut_: this is the last topic so we can bikeshed

alisonmaher: I opened to look for strong opinions
… for naming of priority in AriaNotify
… do we want closer mapping to aria-live?
… I’m of the notion that we should keep separate naming
… any objections?

jcraig: I think `none` and `important` are better than `polite` and `assertive`
… `none` is different than `polite`, so there could be a potential issue there

alisonmaher: yeah, they don’t directly correlate
… `none` is basically saying “not important”
… low priority

jcraig: since priority is optional
… what if we just had an important value
… we currently have an aria-live value of `off` where the announcements don’t occur unless the user is focused on the element
… like a timer

jamesn: should we do some intelligent fallback if someone specifies aria-live of assertive or polite or any unrecognized value

jcraig: a warning in the console can address that

smockle: re: jcraig’s suggestion to collapse priority to a single value, doesn’t that prevent us from adding more granular classifications in the future?

jcraig: not suggesting a single boolean, just a single named value with a fallback to null as a default
… but not a strong preference
… only because `aria-live="off"` was often misinterpreted
… but if consensus is `none`, that’s fine

Matt: in current aria-live, do you have to remove aria-live completely to toggle between the two behaviors?

jcraig: yes, I think you have to remove it

Matt: thinking about user controls where, as a user, it’d be good to have values to represent the 3 possible behaviors
… 1) the element is changing, but you only want it announced if you have focus
… 2) the element is changing, and you want it announced even if you’re not focused

<smockle> s/collapse important/collapse priority/

Matt: 3) the element is changing, and you _don’t_ want it announced

jcraig: for scenario #1, could be solved by having it be a focusable element
… but can’t always anticipate the AT focus
… web author could do that based on activeElement
… the value `none` just seems more similar in semantics to the `off` value than it does the `polite` value
… but alisonmaher might be right that we don’t need the `off` behavior for ariaNotify

alisonmaher: yes, but for verbiage do you think null would be clearer than `none`?

scott: if we were to have another value besides `important`, I’d want it to default to “low”
… `none` sounds similar to `off` and `null` sound like “not applicable”

alisonmaher: yeah, maybe “high priority” and “low priority” are better

smockle: I like high and low
… how about “now” or “later”?

alisonmaher: it’s not guaranteed that “later” won’t actually occur _now_, in reality

Matt: yeah, “how much later?”

jamesn: have we tried ChatGPT? 😉

<sarah> https://www.aiweirdness.com/halloween-candy/

Minutes manually created (not a transcript), formatted by scribe.perl version 240 (Tue Dec 10 03:59:59 2024 UTC).

Diagnostics

Succeeded: s/introduced/interested/

Succeeded: s/scotto/scott/

Succeeded: s/collapse important/collapse priority

Failed: s/collapse important/collapse priority/

Maybe present: BryanGaraventa, dmontalvo, jamesn, jcraig, Matt, scotto

All speakers: Adam_Page, alisonmaher, BryanGaraventa, dmontalvo, filippo-zorzi, giacomo-petri, hdv, jamesn, jcraig, Matt, pkra, sarah, scott, scotto, smockle, spectranaut_

Active on IRC: Adam_Page, alisonmaher, dmontalvo, filippo-zorzi, Francis_Storr, giacomo-petri, hdv, jamesn, katez, pkra, sarah, scott, smockle, spectranaut_