See also: IRC log
<trackbot> Date: 10 February 2014
<richardschwerdtfeger> meeting: W3C WAI-PF ARIA Caucus
<clown> zakim GVoice has Joseph_Scheuhammer
<richardschwerdtfeger> http://lists.w3.org/Archives/Public/public-pfwg/2014Feb/0021.html
<richardschwerdtfeger> scribe: Matt
<mattking> scribe: mattking
Reminder that call time changes on Feb 17 to 1 PM US eastern.
A subgroup working on epub structural semantics starts next week.
JS: is aria 1.1 or 2.0 time frame
<richardschwerdtfeger> http://www.idpf.org/epub/vocab/structure/
Janina: When the module is ready; could be either.
RS: significant amount of work is already completed.
RS: Edupub is a conference. They
are working an extension to what is used in epub for
graphics.
... another WAI taskforce is working on a graphics modeule for
ARIA, probably 2.0 time frame.
<jongunderson> Hopefully they can make the testing sites more accessible too
<jongunderson> We have lots of problems with accommodations with these testing services
Janina: we have 4 votes in
... Encourage reps to vote.
JS: how many are we waiting for.
Michael: the more we get the more it helps with transition. Fewer than 10, could be problematic.
<richardschwerdtfeger> https://www.w3.org/WAI/PF/Group/track/issues/348
<Zakim> jcraig, you wanted to talk about the work involved in deprecating an ARIA 1.0 role name
RS: We want to avoid people using aria-label="" to achieve A11y API omission
JC: Any new role should have a
fall back 1.0 role.
... example: role="none presentation"
... that would work against the goal of having a shorter
token.
Rich previously mentioned that role="presentation" is is a long enough string to effect download times on pages where it is used 1000s of times. So, a shorter string is desirable.
<jcraig> role="<% role %>"
<jcraig> role=""
<jcraig> aria-label=""
JC: role="" as one of the options caries some risk; it could lead to accidental role presentation.
JG: Meaningful role names are
very important to ensure authors understand.
... People read the spec and interpret the spec based on the
name of the role rather by reading the spec.
... I suggest using terms that are meaningful to people based
on their use elsewhere, e.g., inline or block
<clown> <table role="contents" ...
JS: I see the problem is that the
author does not recognize that the content is exposed whereas
semantics is not. The primary issue is one of author
understanding. We need a name that clearly expresses that only
content is exposed.
... So I propose something like role="contents".
RS: Would we entertain some way of having abbreviations.
JC: Using names like block or inline seem more vague because you could have a role block styled as inline, for example.
<richardschwerdtfeger> ach richardschwerdtfeger
Stephan: Role presentation does
not add to understanding because everything is "presented" on a
page.
... There should be some other more clear english term.
<clown> From the spec for the presentation role: "An element whose implicit native role semantics will not be mapped to the accessibility API."
<clown> "The intended use is when an element is used to change the look of the page but does not have all the functional, interactive, or structural relevance implied by the element type"
Stephan: I also do not recommend reusing terms like block, inline, div, etc as that will add to the confusion.
<clown> http://www.w3.org/WAI/PF/aria/roles#presentation
<clown> <table role="layout" ...
s/stephan/stefan/
<clown> <img role="layout" vs. <img role="decorative".
MK: Is one possible approach to narrow the use and then add other roles to cover other use cases.
RS: Originally role presentation was not intended for images, for example.
<Zakim> jcraig, you wanted to mention role="none/null/generic" (the block or inline style could affect final output to API like it does on div/span)
<richardschwerdtfeger> ach richardschwerdtfeger
JC: role="none" or role="null" is
similar in my mind to role="generic", which is a bit
long.
... I have issues with something like block or inline bacause
it ties us to a layout style.
RS: In future, should we consider
whether the layout is actually important.
... and would the term layout be appropriate in other host
language semantics
JC: We still know about position on page even if the role is null or generic.
RS: Because everything in SVG is persistent, we do not want to mapp every element unlessthe element has accessibility semantics. So, in SVG, none could mean leave it out.
JG: Would there be any
restrictions on the use of role="none"
... e.g., would we allow on headings.
JC: No restriction, we are not looking for a new meaning of the role but instead just a syninom
<richardschwerdtfeger> ack richardschwerdtfeger
RS: where does this discussion
leave us:
... Layout is a problem, "" is problematic. Is there is
consensus on the meaning of none?
JG: Can role none only be used to
override native semantics?
... if you can multiple roles, then what would happen if none
is specified as a fallback role?
JC: Maybe that should be an author error.
RS: Doesn't ARIA do exactly that for all roles, override native semantics:
JG? There is a difference between overriding vs removing.
More discussion of need for training to ensure correct usage vs a general tendancy to cause misunderstanding. Even a11y experts have misinterpreted role presentation as meaning that it is left out of the tree completely.
MK: do we need to look at spec to examine the meaning of multiple roles. It is not clear to me thatmultiple roles means fallback and that it is not additive.
JC: The role attribute in ARIA is
an order token list; it is not like a class list.
... For ARIA 1.1, we need to test fall back to ensure that
role="none presentation" results in roole="presentation" to
support legacy aria 1.0 support in browsers.
... Have most objectives to role="none" fallen away.
JS: It is still not my preference, but I do not have a better suggestion.
JD: what should we reconsider how <img role="none" alt="foo"> is handled?
<jcraig> If an element with a role of presentation is focusable, user agents MUST ignore the normal effect of the role and expose the element with implicit native semantics, in order to ensure that the element is both understandable and operable. Authors SHOULD NOT provide meaningful alternative text (for example, use alt="" in HTML4) when the presentation role is applied to an image.
RS: what about <img
role="none" aria-label="foo">
... should be author error.
JC: Spec already calls it out as an error.
<jcraig> <input value="foo" role="none">
RS: nothing in html5 that says alt text must be rendered
<jcraig> <input value="foo" disabled>
<jcraig> <input value="foo" disabled role="text">
<jongunderson> 1+
<clown> http://www.w3.org/TR/wai-aria-1.1/roles#presentation
<scribe> ACTION: jcraig Add role none to spec [recorded in http://www.w3.org/2014/02/10-aria-minutes.html#action01]
<trackbot> Created ACTION-1383 - Add role none to spec [on James Craig - due 2014-02-17].
<jongunderson> bye
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/somethng like role="content"/something like role="contents"/ FAILED: s/stephan/stefan/ Found Scribe: Matt Found Scribe: mattking Inferring ScribeNick: mattking Scribes: Matt, mattking Default Present: Stefan_Schnabel, Michael_Cooper, Janina_Sajka, Rich_Schwerdtfeger, joanie, Matt_King, Joseph_Scheuhammer, Jon_Gunderson, James_Craig Present: Stefan_Schnabel Michael_Cooper Janina_Sajka Rich_Schwerdtfeger joanie Matt_King Joseph_Scheuhammer Jon_Gunderson James_Craig Found Date: 10 Feb 2014 Guessing minutes URL: http://www.w3.org/2014/02/10-aria-minutes.html People with action items: add jcraig none role WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]