– DRAFT –
ARIA WG

New Issue Triage

<chlane> DA has a great voice

New PR Triage

New Issue Triage

<chlane> 1745 ARIA,

<chlane> scotto: wording updates just needs review

<chlane> DA wants to review

<chlane> Rhagav will assign

<chlane> mcooper can set up github

<chlane> accname #160, editorial

<chlane> spectranaut: will look at it

<chlane> aria 1744, editorial

<chlane> turn of linting, more editorial

<chlane> #1743 add files via upload

<MarkMcCarthy> chlane: this happened to me when i tried to upload my change, something to do with github

<MarkMcCarthy> We're merging together PR and Issue triage today due to some growing pains: moving on to issues now

<chlane> back to new issue triage

<chlane> 1748, links children presentational true, flag is missing

<chlane> scotto: 1.4

<chlane> new issue triage, we put them under releases

<chlane> 1747, limit allowed roles aria-haspopup

<chlane> scotto: responded, already limited its use,

<chlane> siri: seeing it used on links

<chlane> on a link or slider, or textbox, what is the purpose?

<chlane> why are we adding it?

<chlane> bryan, it's easier if you have a aria-menu attached

<chlane> spectranaut: contribute to other discussion in 1747

<chlane> and close as dupe

<chlane> core amm 121

<chlane> being worked on

<chlane> 1746, decorative image role

<chlane> scotto: wait until next week, 1.4 for now

<chlane> htmlaam 404

<chlane> further consideration for naming/

<MarkMcCarthy> Issue triage is now done

Deep Dive planning

<chlane> aaronlev: deep dive

<chlane> popup attribute

<chlane> rich tooltips

<chlane> ansynch popup, like a toast

<chlane> floats at top layer

<chlane> working on a11y

<chlane> difficulties

<chlane> if something just pops up is it an alert by default?

<chlane> schedule out a couple of weeks

<chlane> bryan g is out 6/6 - 6/10

<chlane> aaronlev: 6/2?

<siri> +1

<chlane> 9AM 6/2

<chlane> schedule that 9am 6/2

<aaronlev> https://open-ui.org/components/popup.research.explainer

<chlane> self organized chaos

<chlane> scotto: meander through aimlessly

<scotto> ^ appreciate this is in the record

<chlane> lol

<chlane> scotto: lets try to do 6/2

<chlane> will update explainer

Should the ARIA spec suggest setting a description on dialogs?

<chlane> aaronlev: when a dialogs opens usually the first thing that gets focused ... rambling lol, what do screen readers announced when a dialog opens up?

<chlane> every sr has a diff huerisitic

<chlane> some results are not what you want

<chlane> description should be used instead of hueristic

<chlane> spectranaut: put in aria-describedby

<chlane> what would be the spec change?

<chlane> aaronlev: add "ATs should refer to desc of dialog object"

<chlane> scotto: happy to take on

<aaronlev> should prefer description on dialog or alertdialog over a heuristic for reading contents of the dialog

<chlane> question how

<aaronlev> ATs should prefer description on dialog or alertdialog over a heuristic for reading contents of the dialog

<chlane> outside aria scope

<chlane> what is the expectation for native dialogs

<chlane> how do we inform

<chlane> AT how they should handle announcing dialogs

<chlane> aaronlev: its separate

<chlane> try to improve hueristic

<chlane> outside of aria spec

<chlane> siri: why can't we use aria-describedby?

<chlane> screen reader should read whole content

<chlane> programtically focus on heading, so SRs read from top

<chlane> of content

<chlane> trying to understand, what this solves?

<chlane> siri:

<chlane> using aria-describedby, if the author programmatically places focus, what are gaining

<chlane> aaronlev: can use either

<chlane> like labelled or labelledby

<chlane> we want to make sure users get instructions

<chlane> author bypass hueristicv

<chlane> CurtBellew: not having to set focus?

<chlane> aaronlev: focus should go on first tabbable element

<chlane> aaronlev: any useful info that is part of the dialog is read

<chlane> opt in dialog, if you select this such and such poeple will be turned on

<chlane> ok button is focused

<chlane> some SRs will read the ok button only

<chlane> aaronlev: everyone should hear the same thing

<chlane> the spec should say this

<chlane> siri: it answers my question

<chlane> aria-describedby, SRs do specify it, automatic reading is fine, unless there is a lot of content

<chlane> 2-3 paragraphs of text and ok button

<chlane> scotto: the intent is to make these consistently announce

<chlane> siri's situation is a design problem

<chlane> scotto: exceptions can be provided

<chlane> bryan G is in favor of adding

<chlane> CurtBellew: agrees

<chlane> spectranaut: discuss further next time

<chlane> scotto: pushing out scope definition lol

indicate what to do with multiple headings in hgroup

<chlane> aam 398, companion to updated outline algorithm

<chlane> there is no algorithm

<chlane> no underlying magic

<chlane> that tooks away hgroups non incremented powers

<chlane> relook at it for what it could possible do rather than being a generic

<chlane> multiple heading elements

<chlane> single heading followed by a paragraph that represent sub headings

<chlane> do we want to keep it as a generic

<chlane> or is there value to exposing as group or heading group

<chlane> dpub doc subtitle, aligns with heading role,

<chlane> role=subheading

<chlane> AT navigate by subheading someday

<chlane> need to know, do we want to change mapping from generic to something else

<chlane> paragraphs to subtitles

<chlane> scotto: hgroup > h1 > h2 > h3

<chlane> h3's would be downgraded to paragraphs

<chlane> new content model only allows for a single heading and paragraphs

<chlane> ARIA would need to be involve for subheadings

<chlane> jamesn: wait until it shakes out

<chlane> scotto: makes sense

<chlane> jamesn: direction of that will provide answer

<chlane> scott will put in on our radar

<chlane> scotto: wait until it's needed

<chlane> same can be said about mapping for hgroup

<chlane> jamesn: wait to not do it twice, premature

Does aria-hidden obey DOM or AX tree ancestors?

<chlane> aaronlev: it should obey AX tree

<chlane> when it is an accessibility property

<chlane> spectranaut: safari hasn't implemented aria-owns

<chlane> FF doesn't respect hiding children

<chlane> aaronlev: safari aria-owns is the bigger problem

<chlane> jamesn: not anywhere?

<chlane> scotto: yeah

<chlane> jamesn: ouch

<chlane> sarah_higley: popup and of dom with aria-owns doesn't work in ios voiceover

<chlane> aaronlev: aria-owns caused a lot of pain

<chlane> spectranaut: it is on MDN

<chlane> aaronlev: some ARIA works in some places is a pro=blem

<chlane> scotto: site called aria owns you

<chlane> spectranaut: does aria owns obey ax tree

<chlane> jamesn: just hide that thing as well

<chlane> jamesn: 2 choices and one is a non starter

<chlane> aaronlev: it's gotta be using aria-owns

<siri> lol

<chlane> spectranaut: soft objection from jamesn

<chlane> scotto: agree, visible text that is inaccessible

<chlane> scotto: note popup toggletip where content is pulled in for aria-owns

<chlane> the element that visually renders is not

<chlane> aria-owns is a hack

<chlane> if you get burned by it there you go

<chlane> cc jcraig about status of aria-owns in safari and we agree

<chlane> rrsa, make minutes