W3C

– DRAFT –
(MEETING TITLE)

24 March 2022

Attendees

Present
aaronlev, jacobo, Jaunita_George, Jem, joeyang_, MarkMcCarthy, Matt_King, myasonik, scotto, StefanS
Regrets
-
Chair
-
Scribe
chlane

Meeting minutes

New Issue Triage

1714 Does aria-hidden obey DOM or AX tree ancestors?, james milestone this in next v of aria

scott james 1.3 issue

1712

1.4 mileston james

menu item no pres children 1.4?

scott heading inside menuitems

prose for what is allowed as children

buttons and tablists not explictly excluded or implicitly included

allowed child element, need prose

heading should not be one of them

requieed context role is not equivalent to allow the context role

matt/scott

triage

scotto and sarah

milestone for 1.4

1710

Which roles serve as group names and descriptions?

1.4?

scotto: related to accname stuff

jamesn:

more infor

info

112

Change mapping for generic

scotto: div span mappings

assign to matt lol

152 Tooltip def considerations

melsumner: to not block prs broke this out

1707

change events not defined

Specifically call out expectations for dynamically updated descriptions

implementers not aware

1.4

scotto: taking it

Expose a dialog as modal / non-modal #377

111,

Difference between combobox at an input or at another element

valerie volunteered by james

New PR Triage

1713

jamesn: closing

w3c/accname #150

w3c/aria #1705

melsumner: volunteered

volunteed chlane mark

Deep Dive planning

dialogs next week

hidden content - https://github.com/w3c/accname/issues/57

When is hidden content taken into calculation of name and description? #57

spec say only direct children will be exposed

aaronlev: trying to detect doubly hidden stuff is a non starter

if you point to something that is hidden you get all the ax tree content

youll get all the hidden content

expect for a wierd situation

youll get both visible and invisible content

matt, take out words about explicitly hidden

programmatically knowable?

aaronlev: not now, maybe tecnically possible

jamesn: assumes chromium is happy with it?

jacobo: chrome is hiding things that are aria-hidden

matt there was mention of aria hidden

counterintuitive

the purpose is to hide from AT

I'm invisble

in name calc is not correct

jamesn:

jcraig: display none could get shown

cyns: use case where people had labels for screen reader text ariahidden and offescreen

as way to put it in the element without double reading

used to be a commen patter

n

matt, why?

offscreen is visible to accname

display: none will prevent double reading

aaronlev: matt, we went down roads, tooltip, decided, wording easier to have hidden stuff/not hidden stuff

when you point to something that is hidden you get everything

hairy

aria-hidden misuse

still need accessible name

aaronlev: when something gets focus labelledby references are still used

matt chrome bug, does this

aaronlev: will change this

aaronlev: spec hard to understand

matt is aria hidden completely ignorerd?

aaronlev: is subtree hidden? set bit after that we include every node in acc name

jamesn: aria hidden as child, seems useful, allows author to correct behavior

authoring possibilities

spec is already complex, this wont add to that

is this implememtanle by other browser engines

FF follows this behavior

<Zakim> jamesn, you wanted to ask how widely this has been shopped to other browsers

jcraig: seems reasonable

feesible

what can be agreed on an easily implentable

need to do full review, on his list

jamesn: prefers consistency among user agents

matt, we are making an edge case for aria-hidden

except when it is inside a hidden element and referenced by a name calc

scotto: aria-hidden="sometimes"

lo

l

<scotto> i appreciate that's how i was minuted

<jamesn> aria=hidden="🤷🏻‍♂️"

two wrongs dont make a right, should try to simplify

<cyns> +1

matt, we can do a better job of explaining the limitations of aria-hidden

jamesn: accnname is written for implementers

matt we live with the consequences

cyns: canonical test cases

matt bryan did that

with the accname calculator

bryan, test cases archive needs to be revisited

MichaelC: did we evergreen accname pretty sure

yes it is marked as a living standard

siri, hidden referring to visibility hidden and display;none, are you giving extra info to screen reader users now

cyns: this is by design

an important use case

tooltips that don't need the user to diplay them

scotto: repeat tooltip

lol

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

Diagnostics

Maybe present: cyns, display, jamesn, jcraig, melsumner, MichaelC