W3C

– DRAFT –
ARIA WG

14 April 2022

Attendees

Present
CurtBellew, jamesn, Matt_King, myasonik, sarah_higley, scotto, Siri, spectranaut, StefanS
Regrets
HarrisSchneiderman, PeterKrautzberger
Chair
-
Scribe
chlane

Meeting minutes

Working Group news!

New Issue Triage

154 Bryan hasn't had time to see scotto was looking at an old HTML AAM issue

there was a suggestion to cover case where label contains a form control and points to another one

scotto: suggests to call this out as maybe do it but invalid if you do it

jamesn: milestone this for accname

w3c/aria #1721

Consider a mechanism to associate controls without an explicit grouping

1721

scotto: goes back to radio button and related to open UI

specing radio and radiogroup

html allows for associations without radio group

allow for markup patterns that don't neatly fit

jamesn: arent these patterns broken

scotto: arguable

scotto: a lot of discussion allowing for association of other elements, checboxes, toolbar, x of y

StefanS: embed in li

StefanS: when you arrow to toolbar items, described by is not elegant

StefanS: will work on this

scotto: will keep eye on it

w3c/aria #1720

Add explicit value calculations for textbox and searchbox role

nothing in it

jamesn: combobox textbox 1.4

need to say something about getting value

scotto: behaves as expected, describing reality

jamesn: 1.3

New PR Triage

https://bit.ly/3JBcdVR

w3c/aria #1719

Update authoring requirement for aria-selected on options

matt sarah and chlane reviewers

Deep Dive planning

4/28 open UI deep dive scheduled

5/5 dialog deep dive

scheduled

open UI catchup is tentative

dialog deep dive is confirmed 5/5

some weeks we have in depth meeting an hour before this one

TPAC Announcement

TPAC is going to be hybrid

reg will open early July

will be 9/12-9/16

in Vancouver

remote will be better that in the pasty

past

Support aria-description

Support aria-description #69

jamesn: Bryan update?

Bryan, doesn't know the current status, James C and Aaron working on details

Bryan will go back and look

scotto: will review as well

When is hidden content taken into calculation of name and description?

jamesn: waitng for reviews

from James C and Cynthia

Initial aria-textseparation (depends on generic PR being merged)

Matt_King: no progress yet, not prioritized, aria-haspopup was higher

Matt_King: wants to work on it

jamesn: not convinced of pressing need for text separation

Matt_King: popup is higher prioirity

scotto: if we could just add a note to core AAM

scotto: will file an issue

Inconsistency between native and ARIA listboxes when implicit aria-selected is provided

sarah_higley: no need to discuss

jamesn: have solution

Secondary actions on items in composite widget roles

<sarah_higley> https://github.com/w3c/aria/issues/1440#issuecomment-1091984559

sarah_higley: will put the link in the gist too

steps are listed in the comment

sarah_higley: change the wording "authors may use aria-keyshortcuts"

Matt_King: 3rd most permissive scoping, nameable

Matt_King: wondering whether starting with broad scope is a good idea

just becuase of the amount of testing

large amount of examples

to avoid situaitons from the past

don't limit unecessarily

when introducing something new

going big as the start has greater potential for chaos

would we be open to start with a more narrowly scoped option

jamesC is pushing for more permissive scope

jamesn: do we risk locking AT in behaviors around what we have done now

Matt_King: if that risk exists there is a problem

with extending it, period.

Matt_King: if the use cases are that different, it shouldn't be broad

jamesn: why is that?

jamesn: if allowed on a broad set of things, constraints may make the interface decision different

risky to narrowly scope originally

Matt_King: the risk in designing broadly will make it more difficult to get implementation in AT

Matt_King: need conversation with AT folks

tests, colloborate with authors and AT devs

the rest of the world will come

success depends on bringing on all stakeholders

end to end from the start

jamesn: no point in doing the work

scotto: aria readonly on checkboxes

sarah talk to jaws and nvda

Matt_King: has meeting with vispero tomorrow

Matt_King: this is actually, going to add this to the agenda in concept

jamesn: out tomorrow

Matt_King: not tomorrow, next week

Matt_King: not talk about secondary actions specifically

Matt_King: 1 goal get feedback on test

with signoff process

Matt_King: the dream for the project, involve them from beginning in the developoment of tests

this is how we might to ARIA in the futute

future

Matt_King: awesome synergy

Matt_King: it was already hard to get stuff in the spec

this is increasing scope

Matt_King: the more we can turn this into a process and the right people to work on things

we have stuff in 1.0 that is not used

jamesn: much of the role parity stuff will not be used

scotto: we should leverage them

we should help them with feedback

sarah_higley: still make a PR

Working Group news!

we have found a new co-chair

Valerie brings technical knowledge

valerie excited to help get 1.2 out the door

valerie learned a lot

engineered regression test framework for APG

scotto: congrats and sorry

scotto: corry

jamesn: lives in same town as jamesn

alemeda

Add combobox value support for aria#1225

valerie can close it

not blocking combobox changes

jamesn: fewer user agens how do we get through cr

MichaelC: not before tuesday, brainstorm some proposals

jamesn: only 3 user agents

MichaelC: if we were maintainging html org we would have that angst

Do we need a notifications API in ARIA

<jamesn> https://wicg.github.io/aom/notification-api.html

scotto: make things we are using live regions now

better solution, to pipe what you need to a notification API

Matt_King: can browsers implement for live regions

live regions could leverage this API to be better

scotto: should we have this and also live regions

Matt_King: from authoring perspective you can do both

Matt_King: gazillion ways the DOM can change hard to enumerate

all should be owned by user agent and fed to API for screen reader

just as we have the a11y API we need an API for screen readet

need reliable queuing

if two diff live regions use 2 different ways to render

sarah_higley: are you saying we should be pushing a11y apis to change?

browser managed interactivity

Matt_King: narrator should be able to rely on UIA for anything related to notifications

<jamesn> https://www.w3.org/TR/core-aam-1.2/#mapping_events_visibility

narattor should not be looking at hide/show events

jamesn: coreaaam, has events that fire when things happen

like when text is removed

jamesn: not sufficient, this is what we have

now

<StefanS> https://experience.sap.com/fiori-design-web/invisible-message/

StefanS: important to invent messaging framework

q chlane

StefanS: important to offer service

<jamesn> I think every framework has one of those

need guidelines to prevent misuse

<jamesn> https://github.com/adobe/react-spectrum/blob/main/packages/%40react-aria/live-announcer/src/LiveAnnouncer.tsx

jamesn: notifications api is from microsoft

live regions punted to 1.4

will go ahead with or without it

jamesn: notifications API is out of scope for ARIA

Matt_King: rrsagent, make minutes

Minutes manually created (not a transcript), formatted by scribe.perl version 185 (Thu Dec 2 18:51:55 2021 UTC).

Diagnostics

Maybe present: MichaelC