W3C

– DRAFT –
ARIA WG

08 May 2025

Attendees

Present
aardrian, CurtBellew, Francis_Storr, giacomo-petri, katez, keithamus, melsumner, pkra, Rahim, scott, Siri, StefanS
Regrets
-
Chair
-
Scribe
spectranaut_

Meeting minutes

New Issue Triage

w3c/aria#2528

james: we kind of support that on a few things right

<MarioB> join+

scott: we are getting some automated errors with firefox and safari

james: on the table ones, we support it on a parent, instead of only on the cell itself

jamesn: its a good idea if we can get browsers to buy in

<aardrian> +1

jamesn: any one disagree?

(none)

melsumner: do we have a label for "needs interop"?

jamesn: I think this one is more like the spec needs the PR first, before it is an interop issue

scott: not a bad idea, after the PR is made

w3c/html-aam#582

<melsumner> FWIW I was curious about it because I saw this same type of issue on some other repo (where some browsers do a thing but some don't)

Rahim: there was an issue that has resurface regarding the validity idl, basically there is a validity state that applies to html input elements. the idl attribute flags whether the element is in a invalid state.

Rahim: this validity idl has a mapping to aria-invalid

Rahim: there are a couple of similar cases in html-aam, where we have an idl element that is impacted by and ARIA attribute, Anna explained that we (the ARIA working group) needs to use the HTML definition of what is causing that thing to be valid or indeterminate

Rahim: So HTML-AAM should probably be updated

scott: thanks for looking into this Rahim, I was wondering in general, steve made this PR a long time ago, it might make sense to have another section about IDL in the html spec rather than sandwhiching this in with the other attributes

scott: we can write an explainer about idl

Rahim: idl has been my life this last year, happy to keep working on this and that idea

mcking: I have a question about aria-invalid though

mcking: conceptually, is it that aria-invalid should be true when a required field is empty?

mcking: I always thought that was annoying

scott: I think we took care of this in other PRs

mcking: Rahim mentioned that specifically, so I was curious

scott: pretty sure we put stuff in there to prevent aria-invalid from being set until someone interacts with it

<melsumner> I thought aria-invalid was more about the validation rather than required?

<melsumner> Like, the data is there but isn't correct?

Rahim: yeah, not to get too deep, validity is an object that contains multiple flags. we can expose aria-invalid true based on these values

mcking: ok great, great to hear we have more control

w3c/aria#2525

mcking: why is this specifically about single page apps?

<Rahim> ValidityState for HTML input elements (reflected as `validity` attribute in IDL): https://html.spec.whatwg.org/multipage/form-control-infrastructure.html#validitystate

adrian: because that is where they are having trouble controlling focus

jamesn: with the amount of stuff going on this, we don't need to discuss in triage, lets agenda for next week and see how the discussion has gone by then

w3c/aria#2521

pkra: you can assign that to me, I already asked a follow up question

pkra: it's a really old part of the spec

pkra: it's an editorial change

New PR Triage

w3c/aria#2527

jamesn: Process update, we discussed with Wilco and axe core maintainers, just like we file issues against browser, we will flag issues that requires checkers to do something

jamesn: to alert people earlier on

jamesn: we have a number of members in this group that are members of ACT, and our staff contact

https://github.com/w3c/aria/issues/2526\

w3c/aria#2526

rahim: will this require implementation change?

spectranaut_: I'm not sure yet, I haven't looked at firefox or chrome yet

WPT Open PRs

web-platform-tests/wpt#52159

Rahim: I'll take a look

Deep Dive planning

jamesn: I don't think we have discussed this: w3c/html-aam#562

jamesn: we need an anaylsis before we can deep dive. anyone else?

TPAC - will we meet? - 10-14 November 2025, Kobe, Japan.

jamesn: we have a survey about whether we are meeting at TPAC. our presumption is that we are meeting, not asking for commitments, but assuming we have enought interest?

<pkra> -1

<katez> -1

<CurtBellew> +1

<Francis_Storr> -1

<keithamus> +1

<janefulton> 0

<spectranaut_> +1

<jamesn> +1

<aardrian> +1

<giacomo-petri> +1

<melsumner> +1

<Matt_King> +1

<StefanS> +1

jamesn: it looks like enough people are going to try to attend

jamesn: I prefer to meet earlier in the week rather than later

jamesn: specifically not meeting on friday

mcking: yeah friday's are a lot cause

adrian: my only hope is that they are lumped together so I can vacation around it ... :)

jamesn: but so much happens at TPAC all week :D

ARIA IDL updates: enumerated attribute conversions, new "enumerated" wai-aria type, IDL examples

jamesn: lots of discussion on this

jamesn: what do you need from us Rahim?

Rahim: the good news is that I converted it from draft to ready to review

Rahim: Anna is happy, I've addressed Keith's concerns.

Rahim: (explains the PR)

Rahim: I'm reasonably confident that we have resolved all the edge cases and issues, and it needs review, but it is substantially

Rahim: for this change, what it actually means? I created a simulator to show

<jamesn> https://rahimabdi.github.io/aria-idl-simulator/

Rahim: clarifies what the attribute actually means

keithamus: (missed question)

Rahim: (answers but scribe didn't catch)

keithamus: I"m a bit confused, it sounded like when the accessibility engine kicks in, it will change the value?

Rahim: the engine will not change the value, but determine which value to expose to ATs

Rahim: the accessibility engine will use the values in the "missing value default" to determine the value to expose in the tree

Rahim: you need an default that you can't get from the content attribute

Rahim: one of the things that helped me in talking to Anna, Anna is separating a difference between the IDL work and the assistive technology experience

Rahim: how is the developer interacting with the IDL, and what is the browser doing seperately

Rahim: aria-orientation, depending on the role, undefined means different things

keithamus: does this need a deep dive? are there web observable changes, have you enumerated them?

Rahim: I have explained a bit in the simulator

keithamus: I'm more curious if a user of javascript will observe any changes, would a javascript user would change how they handle these values?

Rahim: no they aren't observable

pkra: rahim I saw that you are changing the tables, have you checked that aria.js still works

Rahim: I didn't change any tables, I reused existing classes

Rahim: I added a column, I thought it was ok if it used an existing CSS calss

pkra: its more about the data extraction that is needed

pkra: lets talk about this in the editors meeting

pkra: I'll take a look

cyns: you can put me down for review as well

Rahim: about the tool you can use, I would encourage if you play around with it, to understand what setting the content attribute and javascript does

Rahim: the javascript side does validation

Rahim: this should ground what all of this is about

Rahim: take a look there before reviewing the PR

scott: I have a question about WPT test stuff

scott: about opening bugs on browsers

scott: I thought one of the benefits of creating WPT was that it creates issues on browser

webkit doesn't make auto bugs, but apparently firefox and chrome does

scott: I'm fine doing the bug filing, but these seems like extra busy work

scott: I dread filing these bugs, but tester bother me for information

keithamus: on the standards issue trackers, maybe there

<keithamus> mozilla/standards-positions webkit/standards-positions

spectranaut_: It helps to have issues on browsers linked from the description, because then we know when we can merge a PR

spectranaut_: I need to see if there is implementations or implementation support, which I can see by clicking on the linked bug

spectranaut_: I'll try to write guidance on how to make bugs on browsers

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

Diagnostics

Succeeded: s/flag to/great to/

Succeeded: s/internests/interest/

Succeeded: s/miss value/missing value/

Maybe present: adrian, cyns, james, jamesn, mcking, spectranaut_

All speakers: adrian, cyns, james, jamesn, keithamus, mcking, melsumner, pkra, Rahim, scott, spectranaut_

Active on IRC: aardrian, CurtBellew, cyns, Francis_Storr, giacomo-petri, jamesn, janefulton, katez, keithamus, MarioB, Matt_King, melsumner, pkra, Rahim, scott, Siri, spectranaut_, StefanS