See also: IRC log
<trackbot> Date: 07 April 2014
<clown> http://lists.w3.org/Archives/Public/public-pfwg/2014Apr/0048.html
<jcraig> scribe: jcraig
MC: No major news. Forming a Joint TF with SVG Working Group.
JS: Negotiating details with the two groups.
<Zakim> Joseph_Scheuhammer, you wanted to ask if there is an editors' meeting this coming Wed
MC: Standing agenda item. More to
report next week.
... Covers details about editor's meeting: Every other week 30
minutes at 10AM Pactific
issue-523?
<trackbot> issue-523 -- searchbox/searchfield role which is a subrole of textbox/textfield -- open
<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/523
<MichaelC> close issue-523
<trackbot> Closed issue-523.
<MichaelC> different? https://www.w3.org/WAI/PF/Group/track/issues/590
related ACTION-1342 to ISSUE-523
<clown> http://www.w3.org/2014/03/31-aria-minutes.html
JC: not an easy topic, pros and
cons to all options
... suggest tabling for now. Rich and Steve are not here to
discuss
CS: Would like @aria-labelfor and a label role for ease of reading.
<MichaelC> scribe: MichaelC
JC: issues with generic label element
if we had a label role on a heading
might override heading role
so basically, couldn´t have an element with its own role also a label
MK: that´s why this is a big concern for me with labels
with landmarks can just wrap them
but that´s more awkward with label
how does AT expose?
JC: don´t confuse AT with AAPI
MK: not that, but multiple roles
JC: one pro to label role is a validation tool could flag an unreferenced label
<jcraig> MC: what if there are a lable role that existed, but was not required, as in, role="heading label"
<scribe> scribe: jcraig
MC: iow, labels don't have to use
role="label"
... back to issue at hand. caption elements could be given
label role if no better role used
<MichaelC> Straw poll: who would like a label role of some sort?
MC: still have question of whether in absense of labelledby relationship, whether it labels its parent element
<cyns> +1
<joanie> +1
<mattking> +1
± 1 ;-()
<MichaelC> Converse question: who thinks a label role is more trouble than it´s worth?
<joanie> -1
-1
<clown> 0 (sitting on the fence)
<mattking> Depends on restrictions or limitations
<MichaelC> I read modest support in favour of label role
<MichaelC> JD: like explicit roles
<MichaelC> JC: there´s a label role in GTK?
<MichaelC> JD: yes
<MichaelC> JC: is it generic?
<MichaelC> JD: in GTK, static text and labels still have role of label
<MichaelC> MK: is there always a relationship attribute?
<MichaelC> JD: an object with role label has a pointer to what it applies to
<MichaelC> MC: like label for?
<MichaelC> JC: clarifies that ATK handles this better than GTK
JD: label relationship is author explicit, but may be achieved in a variety of ways
<clown> <span arai-labelfor=...?
MC: Might need to require reverse relationship aria-labelfor from label role to element?
<clown> <span role="label" aria-labelfor="…" ?
MK: what about having labelfor but not label role
CS: MS devs wanted to find all labels in document. Would allow them to enforce coding standard on their own projects
<scribe> scribe: MichaelC
MK: potentially you could have role=label, aria-labeledfor, or something referenced by aria-labelledby
JC: big issue here is it adds complexity to the processing rules
in good world, it´s simple
but what if there´s some circularity in authoring of these pieces
so these proposals have benefits but risks
MC: maybe we don´t want all of these options, just pick one
MK: yes
maybe try to look at UAIG from this perspective, and come back with proposal
JC: there are differences in the name calculation alogrithm between spec and UAIG
MC: only non-normative example
JS: a couple examples are different
As it stands, aria-label and aria-labelledby are used to create a "name" property for an accessible. They do not create an accessible of their own
<jcraig> ACTION: pull text computation section into ReSpec include for both Spec and UAIG [recorded in http://www.w3.org/2014/04/07-aria-minutes.html#action01]
<trackbot> Error finding 'pull'. You can review and register nicknames at <https://www.w3.org/WAI/PF/Group/track/users>.
<Zakim> jamesn, you wanted to say that even if we have role=label is should never be an error if that label isn't actually being used
<jcraig> ACTION: jcraig to pull text computation section into ReSpec include for both Spec and UAIG [recorded in http://www.w3.org/2014/04/07-aria-minutes.html#action02]
<trackbot> Created ACTION-1421 - Pull text computation section into respec include for both spec and uaig [on James Craig - due 2014-04-14].
JN: if we do adopt role=label, shouldn´t be required that authors use it
and not an error to use label without reference
jc: why the latter part?
jn: dynamic updates is one use case
hard to update these references during update
jc: could it be a warning, not an error?
jn: people read warnings as errors
what extra benefit?
jc: there´s a meaningful difference in ATK between static text, and static text that is a label
would want to ensure correct use
jn: already hard for developers to label
introducing more possibilities makes it harder
mk: an error to have label for reference nonexistent?
jc, jn: sometimes in evaluators
<jcraig> MC: also discussing if label role labels its parent
<jcraig> MC: could have undesirable effects
jc: think it´s undesirable, it adds more automatic calculation of the a11y tree
which isn´t impossible, but it´s another thing
for instance, table element mapped to table role only if heuristically determined to be a data table
having spec touch this gets complex, best left to implementations
<Zakim> joanie, you wanted to say it's undesirable
issue-590: https://www.w3.org/2014/04/07-aria-minutes.html#item03
<trackbot> Notes added to issue-590 add caption role.
<Zakim> MichaelC, you wanted to point out the caption example means still putting in a pointer
<jcraig> JD: want to be able to expose unreferenced label text as static text, not label
<jcraig> JD: This would disallow this and force us to expose a potentially incorrect relationship'
<Zakim> mattking, you wanted to ask about mappings
MC: label not labeling its parent sounds desirable from what I hear
but doens´t address the caption thing
you could give caption role=label but still have to have a labeling reference
<jcraig> MK: question on next topic: for caption and figcaption. same in html5?
not much of a gain from that perspective
<jcraig> MK: if you map caption to aria-label, is that problematic?
<jcraig> MK: If we have role="label" in ARIA 1.1, and we don't have @aria-labelfor
<Zakim> joanie, you wanted to Make a clarification about what Michael just said
<jcraig> MK: caption role would be unambiguous on table
<jcraig> scribe: MichaelC
Straw poll: Support for label role labeling its parent?
<jcraig> -1
<jamesn> -1
<mattking> 1-
<joanie> -1
<mattking> -1
<clown> +1 only if the user agent makes the connection to the parent.
<joanie> clown: that is the very thing I'm asking not be done
Strong opposition to this proposal
Straw poll: Support for aria-labelfor type attribute?
<joanie> +1
<mattking> +0.5
<jamesn> +0.25
<clown> +.5
<jcraig> +1 to aria-labelfor (-1 to aria-labelfor)
cyns: explore
Moderate support, need to explore more
Straw poll: Support doing both label role and aria-labelfor?
s/aria-labelled-for/aria-labelfor/G
<mattking> ++1
<clown> +1
<jcraig> +1
<cyns> +1
<jamesn> 0
<joanie> +0.75
(Clarification: +1 = support doing both, -1 = do only one of them)
Reasonably strong support for doing both
<Zakim> jcraig, you wanted to clarify Matt's comment that indicates we'd want to create role that is specific to html:table
jc: point out that the proposal on the table for a caption role would apply both to table caption and figcaption
mk: so applies to more than one kind of parent?
jc: yes
mk: so complexity in defining where it can be used
jc: also proposed for a summary element in a description
if we do mint a role only for table, the genericness of ARIA isn´t supported
mk: so as it stands, you need an ARIA labeling pointer on the caption
jc: not if using <caption> in HTML
host language defines the label arc
but ARIA is more host language independent
we could say <caption> has implicit ARIA semantic of label
jc: also there is an action to add a computedRole property
that explicates what the ARIA role worked out to be
so for a <table> <caption> that could return a ¨label¨ role because of the implicit ARIA semantics
<jcraig> HTMLFigCaptionElement.computedRole() could return "label" which would go further to providing 1:1 role mapping to host langs
Straw poll: Support for a dedicated caption role that labels its parent table?
<jcraig> -1
<mattking> -1
<joanie> +0
<clown> +.5
<jamesn> -1
<cyns> -1
JS: note ATK both have caption and label roles
JD: maybe can address in mapping, by saying if it´s in a table map one way, otherwise the other
Fairly strong anti-support for that straw poll
<clown> http://www.w3.org/WAI/PF/aria-1.1/roles#namecalculation
<clown> UAIG version: http://www.w3.org/WAI/PF/aria-implementation-1.1/#mapping_additional_nd_te
<jcraig> Need to consider role(label) and @aria-labelfor in context of @aria-labelledby and name computation algorithm
<scribe> ACTION: king to develop proposal for label role and aria-labelfor property [recorded in http://www.w3.org/2014/04/07-aria-minutes.html#action03]
<trackbot> Created ACTION-1422 - Develop proposal for label role and aria-labelfor property [on Matthew King - due 2014-04-14].
trackbot, associate action-1422 with issue-590
<trackbot> action-1422 (Develop proposal for label role and aria-labelfor property) associated with issue-590.
<clown> action-638?
<trackbot> action-638 -- Gregory Rosmaita to - compose simple examples for aria-owns and aria-controls based upon prose examples in definitions - 2010-04-30 -- due 2010-05-03 -- CLOSED
<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/638
<clown> action-538?
<trackbot> action-538 -- Michael Cooper to Create candidate HTML Bugzilla entries from PF comments for PF consideration -- due 2009-11-11 -- CLOSED
<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/538
jc: proposal for a ¨generic¨ role for div/span, but also for tables determined to be layout
instead of role=¨presentation¨
mk: danger having everything exposed
mc: we´ll pick this topic up as a big one for next week
This is scribe.perl Revision: 1.138 of Date: 2013-04-25 13:59:11 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/@@@/SVG/ Succeeded: s/Joing/Joint/ Succeeded: s/role="/role="heading label"/ Succeeded: s/ATK/GTK/ Succeeded: s/@@/As it stands, aria-label and aria-labelledby are used to create a "name" property for an accessible. They do not create an accessible of their own/ Succeeded: s/question on next topic: for caption and figcaption. same in html5?/MK: question on next topic: for caption and figcaption. same in html5?/ FAILED: s/aria-labelled-for/aria-labelfor/G Succeeded: s/aria-labelledfor/aria-labelfor/G Succeeded: s/@@/ATK/ Succeeded: s/(-1 to aria-labelfor)/(-1 to aria-labelfor)/ Succeeded: s/Q+ to clarify Matt's comment that indicates we'd want to create role that is specific to html:table/scribeNick: MichaelC/ Succeeded: s/scribeNick: MichaelC/scribe: MichaelC/ Succeeded: s/JC: scribe: MichaelC/scribe: MichaelC/ Succeeded: s/MS devs wanted to find all labels in document/MS devs wanted to find all labels in document. Would allow them to enforce coding standard on their own projects/ Succeeded: s/Would allow them to enforce coding standard on their own projects/scribe: MichaelC/ Succeeded: s/CS: scribe: MichaelC/scribe: MichaelC/ Found Scribe: jcraig Inferring ScribeNick: jcraig Found Scribe: MichaelC Inferring ScribeNick: MichaelC Found Scribe: jcraig Inferring ScribeNick: jcraig Found Scribe: MichaelC Inferring ScribeNick: MichaelC Found Scribe: MichaelC Inferring ScribeNick: MichaelC Scribes: jcraig, MichaelC ScribeNicks: jcraig, MichaelC Default Present: Michael_Cooper, Janina, Bryan_Garaventa, James_Nurthen, Joanie_Diggs, Joseph_Scheuhammer, Matt_King, jcraig, Stefan_Schnabel Present: Michael_Cooper Janina Bryan_Garaventa James_Nurthen Joanie_Diggs Joseph_Scheuhammer Matt_King James_Craig Stefan_Schnabel Cynthia_Shelly Regrets: Richard_Schwerdtfeger Agenda: http://lists.w3.org/Archives/Public/public-pfwg/2014Apr/0048.html Found Date: 07 Apr 2014 Guessing minutes URL: http://www.w3.org/2014/04/07-aria-minutes.html People with action items: jcraig king pull WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]