Minutes: August 6, 2015 WAI-ARIA Caucus

Minutes:
http://www.w3.org/2015/08/06-aria-minutes.html


Text:

Protocols and Formats Working Group Teleconference
06 Aug 2015


Agenda


See also: IRC log


Attendees
Present
      Tzviya_Siegman, Janina, Joanmarie_Diggs, Jemma_Jaeun_Ku, MichaelC,
      jongund, Michiel, Bryan, Bryan_Garaventa, Rich
Regrets
      JamesC, Fred
Chair
      Rich
Scribe
      jongund, joanie, Jemma
Contents
      Topics
         1. aria-describedat meeting
         2. Test planning
         3. TPAC FTF meeting
         4. ICONBUTTON role
         5. Issue 694
         6. Issue 1666: aria-current should be tolerant to user-defined
            values Alex Surkov
      Summary of Action Items



<trackbot> Date: 06 August 2015


<MichaelC> meeting: ARIA


<MichaelC> scribe: jongund


<scribe> scribe: jongund


aria-describedat meeting


TOPIC aria-describedat


JS: Some guests will be coming next week


<tzviya>
https://lists.w3.org/Archives/Public/public-digipub-ig/2015Jul/0042.html



JS: They will be providing some examples before the next meeting
... Some publishers will be joining to describe their requirements
... There are some objections, but we will be following up after the
meeting to dot all the i's and cross the t's
... Educational laws are requiring accessible books


Test planning


MC: After the august 15th meeting we need to talk about testing
... We need to get to CR within the next few months
... We want to have testing well on its way before going to CR, to make
sure there are no suprises
... We plan to continue to use the test harness from ARIA 1.0, need to talk
to system admin to get it setup
... We need to prepare test cases and test files for ARIA 1.1
... Rich wants to make this a top priority after aria-describedat
... Anything else on testing?


TPAC FTF meeting


MC: The URL is in the agenda


<MichaelC> TPAC 2015


<MichaelC> Draft ARIA TPAC planning


MC: The ARIA working group is meeting the first part of the meeting (monday
and tuesday)
... If you are going to just monday and tuesday register under PF


If you are going to just monday and tuesday register under PF for just
monday and tuesday


MC: DIgital publishing might be a topic


JS: HTML5 will be Thursday and Firday


MC: updates the wiki: https://www.w3.org/WAI/PF/wiki/Meetings/TPAC2015/ARIA

... Might be some overlapping meeting conflicts with DPUB


Tzviya: DPUB meeting comment


JS: Graphics SVG


MC: The SVG working group has cancelled their meeting at TPAC


JS: Probably testing


MC: Two topics: agenda items and meeting with other working groups


JS: I think we have done the working groups


MC: Digital publishing web apps
... The TAG is meeting on mondy and tuesday, we can bring grievances to
them


JS: SOme of the WAPPS issues
... The coordinating is getting some time, but we also need agenda topics


MC: Deliverables plan for the upcoming year, the current gant chart is
being developed


Tzviya: DPUB has not schedules time with CSS


MC: I updated wiki pages


https://www.w3.org/WAI/PF/wiki/Meetings/TPAC2015/ARIA



MC: Register for hotels as soon as possible and get your flights


Tzviya: There maybe better hotel rates than the conference rate


MC: Someone was looking for stuff to do at TPAC, some options were sent to
team, I will try to make public
... We need to start looking harder for rich


JS: I will try to contact


RS: TPAC FTF


MC: We have covered that and put some items in the agenda


RS: CSS going to be there?


JS: Thursday and Firday


RS: We need to talk to CSS


JS: Michael can we put that on the wiki page


MC: We just have web apps and digital publishing


RS: Do we know who will attend


Tzviya: Alan will be there


RS: I did not sign up to meet with CSS
... Do I need to resign up for these working groups


JS: Yes indeed, this is not optimal
... You are covered as chiar and facilitator, a very good thing
... Tzviya is also included


RS: Who is coming from PF


Jemma: no


JS: Yes


JRG: no


MC: Yes


RS: We need dominic and someone from apple
... I encourage people to come, exposing a API to applications
... To get where we need to go, there will need to be an API to
interoperate with AT, it could also support alt input devices


ICONBUTTON role


RS: Apparently alot of use cses where there are visual icons on the screen,
people want people to know they are an icon button role
... A visual symbol versus a text button or other symbols on the screen
... The symbols are often standardized, they are very common in SVG drawing


JG: What would you tell an author


RS: I am not convinced


JG: We already have alot of roles in ARIA what is would you tell authors


RS: I am happy to take that back to the group, I wil take it back to the
group


MB: What are the ...


RS: Does anyone object to closing this, having a role of label with an
image


Joanie: What will an assistive technology do with this
... We have a role description property that could describe this
... If an AT has to do something different use a role, if it is more
descriptive use aria-roledescription


RS: Cogntive will have something related to tasks or functions, like a
'Help" buttons, and this would be standardized across applications
... I don't see it for this example


RESOLUTION: We do not believe there is significant difference ....


JS: No meaningful information to the user


RS: Typing in a proposed resolution


JS: At both ends of the chain we do not see a benefit
... RS drafting a response ...


<richardschwerdtfeger> proposal: We think that there is not signficant
additional information to the user when conveyed through accessibility APIs
and in fact treating this role different from a button would likely add
confusion to the user, consequently, we recommend not introducing an
iconbutton role and if SVG A11y Task force requires more detail on the role
they can use the new roledescription attribute which would require
translation


<janina> +1


<Jjemma> +1


<MichielBijl> +1


+1


<joanie> +1


RESOLUTION: proposal: We think that there is not signficant additional
information to the user when conveyed through accessibility APIs and in
fact treating this role different from a button would likely add confusion
to the user, consequently, we recommend not introducing an iconbutton role
and if SVG A11y Task force requires more detail on the role they can use
the new roledescription...
... attribute which would require translation


<richardschwerdtfeger> RESOLUTION: close action 1666


Issue 694


Issue 694 aria-busy description does not limit use to live regions but
characteristics table and UAIG do limit its use


• https://www.w3.org/WAI/PF/Group/track/issues/694



RS: ... reading issue ...
... It sounds like matt would like to simplify the definition


BG: A lot of people do not understand what it is suppose to do


<richardschwerdtfeger>
https://rawgit.com/w3c/aria/master/aria/aria.html#aria-busy



BG: Signals that the live region is complete and start speaking stuff


RS: You can have sub trees as well


BG: the confusion is AT vendors are considering it hidden


RS: Freedom Scientific had a discussion about that a few weeks ago
... do we want to change the text to live region is still being updated


BG: That sounds good


RS: will we lose something?
... we could start out with that, we could add all the elements of the sub
tree


MB: There is valuable info in the current description, multi-part loading,
fails to load is also valuable


RS: OK
... Do you want to change the first sentence?


MB: Could indicates the live region is still being updated


RS: DO we want to say the live region is still being updated ...


MB: Yeah
... The default is false does not make sense


RS: Yes it does ...
... That basically says you can dissect the tree


MB: You set aria-busy=false when the last part is loaded


RS: There is no author must
... Change the first sentence to say....


<richardschwerdtfeger> Proposal: change the first sentence to say:
“"Indicates whether a live region is still being updated”


MB: We could lose the the part after ....
... BG said an entire document is not marked up as false


RS: It can be any part of a web page, it does not have to be a widget


JG: Some people want to use this to create their own user experience


RS: There is a company does this for the login


MB: Maybe I am thrown off by the example ... intialized and complete, page
thew me off


RS: Change this to element
... Do we need this change?


Joanie: I have opinions when it effect orca or questionable, this is
neither to me


JG: SOme people add aria-live regions to menus because they don't
understand ARIA


<richardschwerdtfeger> "Indicates that a live region or widget is being
updated.


<richardschwerdtfeger> The default is that aria-busy is false. If authors
know that multiple parts of the same element need to be loaded or modified,
they can set aria-busy to true when the first part is loaded, and then set
aria-busy to false when the last part is loaded. When a widget is missing
required owned elements due to script execution or loading, authors MUST
mark a containing element with aria-busy equal to true. For example, until
a section of a page is fully


<richardschwerdtfeger> initialized and complete, an author could mark the
document element as busy. If there is an error updating the element, author
MAY set the aria-invalid attribute to true."


<richardschwerdtfeger> The above description is written to apply to any
busy element. Yet, the true value definition is:


<richardschwerdtfeger> "Indicates that a live region or widget is being
updated.


<richardschwerdtfeger> The default is that aria-busy is false. If authors
know that multiple parts of the same element need to be loaded or modified,
they can set aria-busy to true when the first part is loaded, and then set
aria-busy to false when the last part is loaded. When a widget is missing
required owned elements due to script execution or loading, authors MUST
mark a containing element with aria-busy equal to true. For example, until
a section of a page is fully


<richardschwerdtfeger> initialized and complete, an author could mark the
document element as busy. If there is an error updating the element, author
MAY set the aria-invalid attribute to true."


RS: Is it good enough?


MB: It is better than before


<joanie> widget being updated may be problematic as is


MB: We need to wait for her to call back in


GB: I do not use aria-busy, it does not seem to make a difference


RS: That maybe the AT, we don't want to code to the AT


GB: Users only can interact with what's there


RS: What if there are network traffic problems


GB: SOme people put up please wait


RS: i don't see that used in practice, aria-busy is a global attribute
... If you are waiting for a form to load can put it on the form


<MichielBijl> progressbars can also use aria-busy


GB: The UA explains the association between aria-busy and aria-live in the
implementation guide, so there is some confusion to me....


RS: I am looking at perceivable regions, what does it say about ...


GB: It doesn't seem to change any behavior of AT


RS: In the glossary ....


GB: I can send the link later


RS: There is a section on the definition of a live region in the glossary,
reads definition...
... The use cases included wdigets
... You should not make a selection until it is all there


GB: If it is unclear to me, it is probably a problem for other people


RS: If you can say a live region or a widget is still being updated is that
enough


Joanie: Can I just jump back in


<richardschwerdtfeger> Indicates that a live region or widget is being
updated.


<richardschwerdtfeger> The default is that aria-busy is false. If authors
know that multiple parts of the same element need to be loaded or modified,
they can set aria-busy to true when the first part is loaded, and then set
aria-busy to false when the last part is loaded. When a widget is missing
required owned elements due to script execution or loading, authors MUST
mark a containing element with aria-busy equal to true. For example, until
a section of a page is fully


<richardschwerdtfeger> initialized and complete, an author could mark the
document element as busy. If there is an error updating the element, author
MAY set the aria-invalid attribute to true.


Joanie: All the discussion makes sense, my concern is the word "updated",
anything can be happening, we need like populated


RS: I like the word populated better


MB: me too


<joanie> proposal: Indicates that a live region is being updated or a
widget being populated.


RS: the AT can covey the information to the user


GB: I like populated better


RS: I want to make sure it is clear


Joanie: I have a concern about "multiple parts of the same element" ...
... On the same element


RS: James made the suggestion, if the widget is not fully built ....


Joanie: What if the element is a parent element, with two child elements
that are populating, set on parent element


RS: When a widget is missing required owned elements


Joanie: What if it is a form, if the author knows ..... set to true is when
the first list box is starting to be loaded and false when the last listbox
completes loading


MB: aria-busy should be on the form


Joanie: Sometimes it is not required elements


RS: We need to look at this from an authoring perspective


<richardschwerdtfeger> RESOLUTION: Change the first 2 paragraphs of
aria-busy to read as follows: “Indicates that a live region or widget is
being populated.


<richardschwerdtfeger> The default is that aria-busy is false. If authors
know that multiple parts of the same element need to be loaded or modified,
they can set aria-busy to true when the first part is loaded, and then set
aria-busy to false when the last part is loaded. When a widget is missing
required owned elements due to script execution or loading, authors MUST
mark a containing element with aria-busy equal to true. For example, until
a section of a page is fully


<richardschwerdtfeger> initialized and complete, an author could mark the
document element as busy. If there is an error updating the element, author
MAY set the aria-invalid attribute to true.”


Joanie: I thought we were going to let the APG work with this


<richardschwerdtfeger> ACTION: Joanie to Modify the aria-busy text to
reflect the new aria-busy text [recorded in
http://www.w3.org/2015/08/06-aria-minutes.html#action01]


<trackbot> Created ACTION-1699 - Modify the aria-busy text to reflect the
new aria-busy text [on Joanmarie Diggs - due 2015-08-13].


RS: We will ask them to review and suggest changes if they want any
... It is better than what we had


<richardschwerdtfeger> Action-1799?


<trackbot> Sorry, but Action-1799 does not exist.


<joanie> action-1699


<trackbot> action-1699 -- Joanmarie Diggs to Modify the aria-busy text to
reflect the new aria-busy text -- due 2015-08-13 -- OPEN


<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1699



<richardschwerdtfeger> Action-1699?


<trackbot> Action-1699 -- Joanmarie Diggs to Modify the aria-busy text to
reflect the new aria-busy text -- due 2015-08-13 -- OPEN


<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1699



Issue 1666: aria-current should be tolerant to user-defined values Alex
Surkov


<MichielBijl> https://www.w3.org/WAI/PF/Group/track/issues/716



RS: ... reading the issue ...
... Is not aria-current a boolean


<MichielBijl> http://www.w3.org/TR/wai-aria-1.1/#aria-current



Joanie: It started out simple, not there are 5 or 6 different types


GB: It is suppose to support true


MB: ... date, time, true, false...


<richardschwerdtfeger>
https://rawgit.com/w3c/aria/master/aria/aria.html#aria-current



RS: It is currently tokens
... We want to make this translatable, unless we call it user defined


MB: Do we have examples of custom values


RSL I do not this was alex request


Joanie: I wonder if this related performance
... There is a sanity check of these strings


RS: ... validilty check...
... What he wants it to be is a string....
... If we say these are reserved words, these are predefined values, or let
them ....


MB: What is the problem with custom values


RS: ATs don't know what to do with it


MB: Is it a problem for the AT?


RS: If they have not seen it before they .... if something sia step in the
process is valuable
... Date token is in a calendar...., we can put the user as a string and
then it needs to be translated


Joanie: I like currentfor people didn't like that so now we have tokens
... Where I am going with this is if it is any string than what does it do
with it, only need to localize if it is going to speak it
... What is an AT support to do with this?


RS: If it is a page in a list, and you have a page number, ie I am this far
into a document
... Bread crumbs is an example of this


Joanie: orca can look at this and say the current page or current step
... These values are programmitically useful


RS: A separate attribute for user translatable...


<joanie> scribe: joanie


<Jjemma> scribe: Jemma


<joanie> scribenick: Jjemma


Joanie: the problem is there is no context for AT, ie. when AT will speak.


RS: no semantics for this aria current


joanie: is not clear where user is at, child or parent component when it
says "current"


<richardschwerdtfeger> Proposal: The WAI-ARIA spec. says that user agents
SHOULD, NOT MUST, map unknown values to true. The actual value test, for
validity, can be passed to accessibility test tools. Additionally, the
group feels that unknown values are not programmatically processable by ATs
such as notifying the user how far they are along in a breadcrumb list.
Additionally, any custom values would require translattion by the author
per language making custom values less


<richardschwerdtfeger> valuable to the author and assistive technnology.
The group feels that providing custom string values would harm the user
experience, consequently we recommond closing the issue.


<joanie> Joanie: I question what will happen if we say the above, because
it seems like we're giving permission to not validate the values.


<joanie> Joanie: As a result, I anticipate seeing new and unexpected values
for the object attribute associated with aria-current.


<joanie> Joanie: In which case, why bother suggesting validation?


<richardschwerdtfeger> Proposal: The aria-current attribute is an
enumerated type. Any value not included in the list of allowed values
SHOULD be treated by assistive technologies as if the value true had been
provided. If the attribute is not present or its value is an empty string,
the default value of false applies and the aria-current state MUST NOT be
exposed by user agents or assistive technologies.


<MichielBijl> +1


<richardschwerdtfeger> +1


<joanie> +1


<janina> +1


+1


<richardschwerdtfeger> RESOLUTION: change the aria-current text second
paragraph to say: The aria-current attribute is an enumerated type. Any
value not included in the list of allowed values SHOULD be treated by user
agents or assistive technologies as if the value true had been provided. If
the attribute is not present or its value is an empty string, the default
value of false applies and the aria-current state MUST NOT be exposed by
user agents or assistive technol[CUT]


<richardschwerdtfeger> ACTION: Joanie: Modify the ARIA spec. to reflect the
changed text of the second pagraph for aria-current [recorded in
http://www.w3.org/2015/08/06-aria-minutes.html#action02]


<trackbot> Created ACTION-1700 - Modify the aria spec. to reflect the
changed text of the second pagraph for aria-current [on Joanmarie Diggs -
due 2015-08-13].


<joanie> trackbot, associate action-1700 with issue-716


<trackbot> action-1700 (Modify the aria spec. to reflect the changed text
of the second pagraph for aria-current) associated with issue-716.


Summary of Action Items
[NEW] ACTION: Joanie to Modify the aria-busy text to reflect the new
aria-busy text [recorded in
http://www.w3.org/2015/08/06-aria-minutes.html#action01]
[NEW] ACTION: Joanie: Modify the ARIA spec. to reflect the changed text of
the second pagraph for aria-current [recorded in
http://www.w3.org/2015/08/06-aria-minutes.html#action02]

http://www.w3.org/2015/08/06-aria-minutes.html

Rich Schwerdtfeger