<jamesn> https://react-spectrum.adobe.com/
<scribe> scribe: Matt_King
No new issues!!!
jn: oo, surprised, is somethingwrong/
Issue 1453
oops, that is a practices issue.
First issue is for aria editors.
jn: Michael can you put it on editors agenda.
pr 1298
<jamesn> https://github.com/w3c/aria/pull/1298
jg: Basically, think example is
overly complicated
... mixing accname and acc desc is confusing people.
jn: reviewers?
scott: Joanie and I worked on including that example dues to confusion related to html caption.
got that example to show that if someone puts in a full paragraph as a caption that perhaps some of that should be part of a description.
jg: think more complex things like that should go in apg
scott: fine by me
Joanie: If authors put in a giant caption and it makes a giant name, then hopefully users will put a stop to it via feedback. Not sure we can fight this in the spec or apg.
JN: maybe put a note to say to avoid long captions to avoid excessively long names
joanie and scott: good with a note
scott: not opposed to change, just providing rationale for why it was what it was
carolyn: add me
jn: should we have a deep dive next week?
carolyn: Woud like to talk about Mac and Windows having different use cases for space key
raised by james craig in practices.
Still working out what the topic is exactly.
jc: will be out for a few weeks
back week of aug 10.
Carolyn: we can wait till James is back
jn: please file issue to generate discussion
Hearing nothing for next week.
If you have something, please file issue before tuesday. as of now, no deep dive meeting.
jn: Don't necessarily look at this suggestion as a way to resolve it. This issue is to generate discussion.
<jamesn> https://github.com/nvaccess/nvda/issues/9187
NVDA announces role switch as pressed checked
<jamesn> https://github.com/nvaccess/nvda/issues/9187#issuecomment-457001861
James Teh comment onjan 23 is inspiration.
James Teh has a suggestion that would expose checked and not expose pressed.
JC: Is there a reason that don't want to expose it as an on/off switch
scott: they have not done that before. it has been expossed, switch has been exposed as a toggle button
jc: surprised nvdea would not want to include that as well.
jn: in core-aam, it gets exposed as a toggle with an object attribute of switch.
Sina: what does that mean, that is exposed as an obj attrib?
jn: there is no native mapping, it is only in object attribute.
Sina: what does that mean to expose it as a switch?
jd: In orca, if it is a switch, I don't say toggle button, I say switch.
If pressed is true, I have Orca announce "on".
JC: Yes, rather than meaning exposed in the AX API, I was talking about how it is conveyed to the user.
Sina: Could that result in a toggle being announced as "On"?
JC: that is a screen reader
implementation decision.
... What is distinction between native role and aria role? If
there is an IA2 property, it is a NVDA choice if they want to
expose that.
Sina: I don't know if I got that from Jamie's comment.
jn: seems like NVDA is exposing exactly what is in the API. they don't have any extra logic here that uses the obj attribs.
Can we do anything to help them? or do they need to resolve this on their own.
jc: Is Jamie talking about the mapping spec, not the aria spec?
jn: I think so
Sina; wondering where the disconnect is. The mapping is to toggle button, not switch.
Joanie, is going above and beyond and renaming the role and the state based on the obj attributes.
is that a mapping issue, or is it a screen reader implementation decision.
jd: Problem is a API gap in IA2.
we could add switch role to IA2.
If they don't want to add a switch role, which is hard, then they should check the obj attribs when rendering toggles.
jn: why did we map to toggle instead of checkbox?
carolyn: maybe b/c of mixed state
switch cannot be mixed.
jn: but we used the checked state
on the switch.
... it might work be better if switch were mapped to checkbox
and checked.
mk: what is the IA2 state for a toggle button?
carolyn: you suggesting we rewrite the spec
jn: no
might need a core-aam change
scott: seems like the way he resolves this is by ignoring the spec and making switches into checkboxes
Sina: raises author confusion that can happen by putting checked on a button with a switch role.
jn: we could allow either pressed or checked on a button
<Zakim> jamesn, you wanted to try to wrap this
jn: 2 potential issues: 1) do we want to do something to help authors get it right? 2) Do we want to do something in core-aam
seem reasonable
jd: yes
jn: can you log core-aam issue?
jd: yes
sina: is this ever worth resolving by adding aria-on property?
seems like this is not sustainable
to re-use checked
Seems like we do thing that make things less clear than they should be
<joanie> https://github.com/w3c/core-aam/issues/75 is the new issue
<jamesn> thanks joanie
Can we have a generic aria binary property? seems like there is a need for that.
jn: not sure how to turn that
into an action at this time.
... do people support doing this? where would it live? does
someone want to start working on it?
jc: not willing to start it myself but something we should do eventually.
jn: something that would be part of 1.3? or separate?
jc: no need to tie to 1.3. could be separate.
I think this from an accessibility discussion at google.
Maybe a note, maybe not normative
jn: Should we add a comment that says this seems like a good idea but we don't have anyone to work on it. WG is willing to review and advise if someone can start it.
Carolyn: sounds good.
<jamesn> https://github.com/w3c/aria/projects
jn: I put a bunch of projects together. Each is a potential focus area.
Tables and grids have a bunch of issues.
Live region issues is another.
aria-controls is a potential topic.
IDL is another.
attribute parity is a possible project. do we care about that?
jc: It will eventually be something we need for testing.
carolyn: huge effort.
jn: Is the testing issue an aria wg problem or a browser implementer problem.
jc: fair question
Not thinking it is important in the short term.
carolyn: agree that is not a short term priority. We have other things that are broken and need to be fixed now.
jn: OK, decision, attribute parity deprioritized for 1.3.
Braille support is another project, continuing to build on work we have started.
Remaining role parity. how important is that.
jc: what is left.
scott: associationlist is something that needs to be figured out.
jn: aren't they in for 1.3?
scott: yes, want to make sure that lands.
jn: we can keep that going.
But, are there more things left to do that are solveable now?
I'd not object to some kind of prefix approach.
for things that don't already have mappings, e.g., native-___
jc: Could make it clear for web dev tools to give a warning if authors use
jn: We have 1.3 project for role parity, and there is an issue to consider a native or host role prefix for audio, video, etc.
We could extend that issue to include other things that have not yet been mapped and that could wrap all role parity.
mk: would like to take up text separation
jn: yes, hear you, we should keep
that.
... 3 others in hi pri, list type for parity with ol and ul,
and input type for role parity with html input types
the input one could not be solved with a prefix. native-input would not create tneeded differentiation.
This is scribe.perl Revision of Date Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00) Succeeded: s/surprixed/surprised/ Succeeded: s/somethingn /something/ Succeeded: s/fiar/fair/ Default Present: jamesn, Joanmarie_Diggs, irfan, janina, MichaelC, Scott_O, jongund, harris, MarkMccarthy, StefanSchnabel, carmacleod, Matt_King Present: jamesn Joanmarie_Diggs irfan janina MichaelC Scott_O jongund harris MarkMccarthy StefanSchnabel carmacleod Matt_King Found Scribe: Matt_King Inferring ScribeNick: Matt_King WARNING: No date found! Assuming today. (Hint: Specify the W3C IRC log URL, and the date will be determined from that.) Or specify the date like this: <dbooth> Date: 12 Sep 2002 People with action items: WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option. WARNING: IRC log location not specified! (You can ignore this warning if you do not want the generated minutes to contain a link to the original IRC log.)[End of scribe.perl diagnostic output]