<joanie> agenda: this
<joanie> agenda: be done
<scribe> scribe: melanierichards
<joanie> AccName trims whitespace but doesn't define which code points are whitespace #55
joanie: we have talked about doing whitespace as part of 1.2. James, agree?
jamesn: yes we did
<joanie> Add explicit language regarding name from legend and encapsulation #54
joanie: already on 1.2
... re #54, we know what we need to do for acc name calc from
ARIA spec, but it's not actually in accname spec yet
<joanie> separator role splits list in Firefox #49
<joanie> https://github.com/w3c/core-aam/issues/49
joanie: even though it's
core-aam, would like authoring practices people to look at this
issue
... look specifically at steve's test case. There are
separators in menus, the calculation that Gecko is doing, it
includes the separator, and it's also recalculating the groups.
What FF is doing is different from Chromium and maybe also
Webkit. Group needs to decide what's correct and then put it in
the spec
mck: we talked about this a
couple years ago, there are differences in opinion but I don't
know what the answer is here.
... we have a lot of inconsistencies on whether and how
posinset is calculated
<joanie> https://github.com/w3c/aria/issues/1010
joanie: phonetic cues
... APA has a task force for this. I'm thinking close it as not
ARIA?
mck: yeah
... APA is supposed to come with whether it is ARIA, right?
joanie: yes
<joanie> `aria-details` precedence over `aria-describedby` #1009
mck: can we put this on the TPAC
agenda and on 1.2?
... we can't write some ARIA practices because nobody has a
clue what the relationship between details and escriptions
is
jamesn: we're going to need to be able to use both of them
mck: we need browser and AT
people involved in this
... and digital publishing
jongund: we also need a mapping for VoiceOver I think
jamesn: we'll probably have an annotations draft come back to the WG in a week or two
<joanie> https://github.com/w3c/aria/issues/1008
<joanie> Should aria-selected be allowed on row in table?
mck: definitely not
joanie: 1.2?
<joanie> https://github.com/w3c/aria/issues/1007
<joanie> Why can not the author use aria-labelledby if the interface can not display the label on the screen? #1007
joanie: re 1007, some platforms
all have API that if there's that relationship, tries to point
to targets
... [asked Melanie to comment re UIA]
melanierichards: I'll check that out
joanie: nothing to do with name calculation, this is about relationship properties in APIs. What happens if it points to a null reference
<joanie> https://github.com/w3c/aria/pull/1006
<joanie> Remove advice against changing roles
jamesn: already on the agenda
joanie: just your reminder that we need an agenda
<joanie> https://github.com/w3c/aria/labels/F2FCandidate
<joanie> https://github.com/w3c/aria/labels/F2F
joanie: F2FCandidate label is the way you add candidates for the F2F, the "F2F" label is for those topics already approved
jamesn: you can add topics even
if you don't plan to be at TPAC
... just give us direction on what outcomes you're looking
for
joanie: we really need another WD
out the door for ARIA. There are two wiki pages (tinyurls) in
the agenda. Can authoring practices people do a find in page
for "authoring practices:todo"
... if there's nothing to do, change "todo" to "N/A"
... we're missing authoring practices
... non-editorial and other changes
... I think some don't need authoring practices but I need you
guys to tell me
<Jemma> jemma: I think I can work with Matt on this this Friday.
<Jemma> jemma: but James' help is appreicated
mck: we do have a draft for aria-expanded. Some of these we're in a good place, some we're in a bad place. We have a target section for aria-rowindextext etc
<Jemma> Jemma:James, are you mentioning that matt and I, APG, example url, issue url and design spec url to the wiki?
joanie: if it's done, change "to
do" to "done". If not applicable, change to "N/A"
... I just need to know what the status is
jemma: so we're just editing the wiki for status?
mck: yes
<joanie> regret+ Scott
[decided to punt as we're missing some people]
<MarkMccarthy> regret+ pkra
joanie: we need this done for 1.2, can you tackle?
BGaraventa: I can, but I'll be gone all next week
joanie: this to me seems like the
#1 priority regarding the acc name spec
... while I think we can guess at the language, we don't
actually have it in the spec
BGaraventa: on the top of my to-do list
jongund: made the changes awhile ago, is there any more work that needs to be done?
joanie: I see "changes requested", not sure if the changes were made and the flag wasn't cleared. Do you think you addressed them all?
jongund: yeah
joanie: do we want to give people a deadline to review?
jamesn: I think we've had enough time to review
<Jemma> https://github.com/w3c/aria/pull/951/commits/6c643742c2dd8a5ed5584d9e8a56e854cbdce01e
joanie: ok, then we can merge it
<Jemma> jon's last commit
jamesn: unless anyone in this meeting wants to shout
[no shouting]
joanie: after this meeting I will
merge it
... and update the wiki page
... congratulations Jon! That was not an easy one
joanie: if they are empty, they
probably shouldn't have an accessible name
... Harris on the call?
[is not]
jamesn: per his comment, validators are essentially ignoring what we put in the spec
<joanie> https://github.com/w3c/aria/pull/1004
joanie: any reason we shouldn't
merge the PR?
... makes sense to me
jamesn: cell already didn't have acc name required
<Jemma> no objection
joanie: objections to merging?
mck: no
... we should probably say that an empty cell is valid
content
joanie: conclusion is I'm going to merge after the call. Do we need authoring practices on that?
jamesn: what's it say for gridcell?
<Jemma> https://w3c.github.io/aria-practices/#naming_techniques
mck: required only if content is not sufficient. I can probably make an edit to make this even more clear
jamesn: awesome, but we don't require to go forward with the change
mck: no, we don't
joanie: in terms of implementations, what changes?
jamesn: nothing, just
validators
... and at least one already doesn't raise an error on this
jamesn: there is text that
essentially says "don't change the role", but from convos with
implementers, if you do change role, they delete the object and
re-create it. This essentially came out of stuff happening in
WhatWG re custom elements, where they want to be able to do
this, change roles
... I created a PR that removes this advice
<joanie> https://github.com/w3c/aria/pull/1006
<jamesn> https://pr-preview.s3.amazonaws.com/w3c/aria/1006/55a66bc...9873762.html
jamesn: do you think the UA statement should be a SHOULD or a MUST?
<jamesn> https://pr-preview.s3.amazonaws.com/w3c/aria/pull/1006.html#introroles
<Jemma> Should the above SHOULD be a MUST or are there places where this isn't required, for example if I were to change a paragraph to a blockquote - "If a role changes user agents MUST delete the associated mapping and its children and replace it with a new mapping with the appropriate role."
joanie: I think we should remove all this and if we need something for UAs we put it in Core-AAM
jamesn: strike the entire paragraph and editor's note?
joanie: some of it is technically
inaccurate, destruction is kind of a pain...
... the AT is in an object, the UA destroys the object, now the
thing is dead...assuming ATs are handling that scenario, they
will continue handling that scenario
<jamesn> https://pr-preview.s3.amazonaws.com/w3c/aria/pull/1006.html#roles
jamesn: there's some stuff in section 5 we should probably also strike
joanie: if UAs/ATs want a role change notification, belongs in the Core-AAM. If they want the node destroyed, that should go in the Core-AAM. Whatever the user agent is doing is probably fine
mck: so we're saying the ARIA spec shouldn't give any guidance to UAs about what they should do if an author changes the role?
jaonie: if it's plat-specific, should go in Core-AAM
MichaelC: I agree should be in Core-AAM but maybe ARIA spec should say "be extra aware, go check the Core-AAM"
<joanie> https://w3c.github.io/core-aam/#mapping_events_visibility
mck: I think the goal is to reduce Core-AAM to a table of mappings
joanie: right, but this is a
table of mappings
... ATK, UIA, etc probably have different things
mck: so for some you would document as "no mapping"?
joanie: or give the advice to
destroy the object and recreate it
... I will file an issue against Core-AAM
... language will likely have a SHOULD, implementations are
destroying and recreating objects in all platforms
... what's the deadline to merge?
... does anyone disagree with this conclusion?
mck: support it
jemma: so you're going to file an issue on Core-AAM, and reference Core-AAM in ARIA?
joanie: I hadn't planned to but
we could consider going back and doing that
... if we hear from the platform owners that destroying the
node isn't right, we'll have to make implementation changes and
test changes
joanie: we don't have time in the
agenda today, but very important topic
... I suggest we punt today
This is scribe.perl Revision: 1.154 of Date: 2018/09/25 16:35:56 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/[missed]/VoiceOver I think/ Succeeded: s/add url/example url, issue url/ Succeeded: s/objectioni/objection/ Found embedded ScribeOptions: -final *** RESTARTING DUE TO EMBEDDED OPTIONS *** Present: MichaelC Stefan melanierichards Jemma jamesn MarkMccarthy Bryan_Garaventa MattKing (mck) joanie Regrets: Peter_Krautzberger Found Scribe: melanierichards Inferring ScribeNick: melanierichards Found Date: 11 Jul 2019 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]