W3C

– DRAFT –
ARIA WG

03 February 2022

Attendees

Present
BryanGaraventa, carmacleod, jamesn, Jem, Jemma, MarkMcCarthy, pkra, sarah_higley, scotto, StefanS, valerie young, zheng_xu
Regrets
CurtBellew JoanieDiggs, Joanie
Chair
JamesNurthen
Scribe
chlane

Meeting minutes

<jamesn> https://www.w3.org/events/meetings/1d223877-730c-446f-bfa9-2e9222902112/20220203T130000

<jamesn> https://bit.ly/3ARvtLJ

should i run it now?

or wait

#1689 assigned to 1.4

New PR Triage

<jamesn> https://bit.ly/3s6LYzD

<jamesn> https://github.com/w3c/core-aam/pull/107

#107 ready to merge no disagreements

Deep Dive planning

<jamesn> https://bit.ly/aria-meaty-topic-candidates

no deep dive chosen , no dd on 2/24 may or may not be a meeting will need a chair

Figure out what to do with the DPUB mappings overlap with the ARIA Core spec, in the context of newer implementation differences

<jamesn> https://github.com/w3c/aria/issues/1643

James Craig added updates, James should probabably wait for dpub triage before we do anything else

Is this proposal overcome by events?: Updated spec text to reflect the processing of hidden elements when c…

<jamesn> https://github.com/w3c/accname/pull/137

<pkra> https://github.com/w3c/accname/issues/57

<pkra> https://github.com/jaragunde/accname/pull/1

alex impossible to hide sub sub trees, so once you point to a hidden subtree all nodes would be read

aria-hidden could mean don't use that part of the subtree

they are leaning towards just removing those nodes with JS

clarity for spec is more important

James should close #137 as a non starter

<Jem> as far as it is simper and easy to understand, I strongly support!

Bryan is happy as long as there is agreement

Any updates?: Prohibit rowcount and rowindex on rows in treegrid and prohibit posinset and setsize on rows in tables and grids

<jamesn> https://github.com/w3c/aria/issues/1303

<Jem> https://github.com/w3c/aria-practices/issues/1442

remove it from agenda we agreed

we are going to focus on secondary actions first

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

Do we have any decisions?: Clarify usage of aria-haspopup

<jamesn> https://github.com/w3c/aria/issues/1024

Scott, our last talk he put notes on the 2 directions

returning to true/false or true/false/menu/generic name

more discussion has occcured on the thread since

why keep the attribute?

keep it for backwards compatibility

knowing something is going to popup is not important on the invoking element

bryan, we do needed for menutitem

to indicate subroles

Scott the attribute doesn't make sense on any other element

stefan, dialog grids, why should these be declared?

Scott, there could be an endless number of UI token values

describing everything that could be popped up is untenable

stefan, aria-haspopup could take and ID ref value

Scott, element might not be in the Dom breaking the Id ref

James if there is a visual indication that something will popup

shouldn't we have something in ARIA for that purpose

Bryan is in favor

<Jem> if there is no visual indication of control, it will be the violation of wcag 2.2 visible control.

bryan finds a lot of controls with attached menus

he believes it should support attached interactive roles

if you don't have it, users will not know anything is present

siri aria-haspopup should indicate what is being opened

scott disagrees

Matt wonders why he needs information that others don't get

Scott, do we think there is value informing users somehting has popped up at the control level

Scott, front load that info or let them know what will occurt

James thinks the interction should be defined by the thing that was opened rather then the trigger

bryan, the menu button changes behavior

sighted user expects menu style focus managment

after activating a menu button

move focus to the first menu item

but when that doesn't happen, if they have to press the down arrow for the first item

aria-haspopup is needed for jaws

james mentioned nvda which doesn't do that

matt, there is a role for menubuton in windows

in aria 1.0 we used a property rather than a role

James if all menubuttons were announced as buttons

cyn and james windows users may be confused

james macos reads as popup button

<Jem> https://www.w3.org/TR/wai-aria-practices-1.2/examples/menu-button/menu-button-links.html

matt html select on mac feels like menu

but on windows if feels like a combobox

james/matt macos windows different select behaviors

jcraig: deep dive topic somewhat related to this discussion of ARIA's rigidity, Brian Kardell from Igalia wanted to discuss mixin roles, effectively additive roles (something can be both a "link" and a "heading") which would match the iOS Accessibility API pretty well... any combination of traits.

altenative to ARIA

Matt, mentioned control pattens in UIA

similar to iO

iO

s

matt, decision? deprecate specific values

James specific values don't do any good and are not supported

stefan, why were they invented?

due to combobox 1.1

Bryan, Rich was in favor of token system

James, PR to remove values

Scott, does not solve the issue of notifiying users of rendered UI

<Jem> +1 to scott

and it will dismiss on you if you leave it

<jcraig> [accidental edit]

James overloading popup is a bad idea

matt, make a separate issue

scott, anything can be a popup

people are implementing things that dismmiss in different ways

stefan, how should screen readers announce it?

matt stay with historical precedent

add a menubutton role and get rid of aria-haspopu

stefan it is needed , many wierd designs

james rejects removing aria-haspopup, mentions popovers speech bubble UI

they don't limit popups to menus

James, sighted users know something is going to show up via discovery

james chevrons in zoom could mean popup

but there are no visual distinction

matt, forum in ARIA At

james , can we get a PR that puts aria-haspopup in a better state

Scott and Matt will work on it

Carolyn, true/false/menu if the value outside those values is true

but dialogs should be false

listbox,tree.grid, not used

<Jem> I found this aria-popup discussion is very helpful. thanks, everyone.

bryan is in favor of combobox evaluating to true

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

Diagnostics

Succeeded: s/somewhat related to this discussion of ARIA's rigidity, Brian Kardell from Igalia wanted to discuss mixin roles, effectively additive roles (something can be both a "link" and a "heading") which would match the iOS Accessibility API pretty well... any combination of traits./[accidental edit]/

Succeeded: s/James, deep dive, Bryan JS discuss mixins, additive roles, something can be a "link and a heading"/jcraig: deep dive topic somewhat related to this discussion of ARIA's rigidity, Brian Kardell from Igalia wanted to discuss mixin roles, effectively additive roles (something can be both a "link" and a "heading") which would match the iOS Accessibility API pretty well... any combination of traits.

Maybe present: jcraig