ARIA Working Group -- 03 Mar 2016

<janina> pesent+ Janina

<Rich> https://lists.w3.org/Archives/Public/public-aria/2016Feb/0466.html

<MichielBijl> scribe: MichielBijl

Combobox

not combobox

<Rich> https://lists.w3.org/Archives/Public/public-aria/2016Feb/0466.html

action-2021

<trackbot> action-2021 -- Joanmarie Diggs to Draft text for default max and min spinbutton values -- due 2016-02-18 -- PENDINGREVIEW

<trackbot> http://www.w3.org/WAI/ARIA/track/actions/2021

Action 2021

<Rich> https://rawgit.com/w3c/aria/action-2021/aria/aria.html#spinbutton

RS: min/max negative values?
... What is changed?

JD: min/max size maybe
... Exposure does not belong in spec

RS: Does the author know different values based on platform?

JD: Author doesn't know

<joanie> Authors must set the aria-valuenow attribute. Authors should set the aria-valuemin attribute when its value is greater than negative infinity, and the aria-valuemax attribute when its value is less than positive infinity. If missing or not a number, the implicit values of these attributes are as follows:

<joanie> The implicit value of aria-valuemin is negative infinity.

<joanie> The implicit value of aria-valuemax is positive infinity.

<joanie> The implicit value of aria-valuenow is 0.

RS: Have some text “if you leave them out, UA will set them”

AB: Have not setting values default to infinity sounds logical to me

JD: I've had two shots at this and have not maybe everyone happy

<clown> +1 to joanie's proposal

RS: I'm happy
... joanie you said the HTML spec doesn't mention it?

JS: they use undefined

AB: Whether you want to require a numeric range
... It will always be limited by how the implementation is storing the number
... Can you asure authors that you can store 5.124.131.123?

JD: My spec says that if it's less than infinite the author needs to specify it.

<janina> suggest "value is infinite, whether positive or negative."

<jamesn> how about this? "Authors must set the aria-valuenow attribute. Authors should set the aria-valuemin attribute when there is a minimum value, and the aria-valuemax attribute when there is a maximum value. If missing or not a number, the implicit values of these attributes are as follows:"

<jamesn> "The implicit value of aria-valuemin is that there is no minimum value"

all: that's fine

JD: Does that mean there is no implicit value?

JN: In reality it's the same, no?
... You're not going to write infinite numbers into the API
... Hope not

JS: Math libraries have done this

*infinite discussion about infinity*

RS: I don't have a problem with James' text

<janina> +1

<Rich> +1

JD: So are we using James' text?

<jamesn> +1

<fesch> +1

RESOLUTION: Take James' proposal to modify Joanies proposal.

action-1489

<trackbot> action-1489 -- Michael Cooper to Propose spec text to limit what aria attributes can be overridden by strong native semantics (e.g., aria-label and aria-labelledby) -- due 2016-02-25 -- PENDINGREVIEW

<trackbot> http://www.w3.org/WAI/ARIA/track/actions/1489

topic Action 1489

Action 1489

<clown> action-1489

<trackbot> http://www.w3.org/WAI/ARIA/track/actions/1489

JS: Host language can do preemptive changes

<Rich> https://github.com/w3c/aria/commit/0b9ebc94afa85fcfb281bab88b9dacd2a88111c7

MC: Just added note to end of section

RS: If it's non-normative it can just go in

JS: Looked at original issue
... From James Craig, wants to create a whitelist
... That can't be overwritten

<clown> "Host Conflict Semantics potentially allow host languages to overwrite @aria-label, which could break text alternative computation. ARIA 1.1 should disallow host languages from declaring @aria-label* (and aria-described*?) from being declared in direct semantic conflict with a host language attribute."

RS: Get back to it next week Michael?

MC: We'll see next week

Action 1730

action-1730

<trackbot> action-1730 -- Michael Cooper to Draft ¨deprecated¨ section for conformance section of spec -- due 2015-10-14 -- PENDINGREVIEW

<trackbot> http://www.w3.org/WAI/ARIA/track/actions/1730

<Rich> https://rawgit.com/w3c/aria/ACTION-1730/aria/aria.html#deprecated

MC: not needed? there is nothing that says it's deprecated.

RS: Do you think we have to put that in the list of terms?

MC: Maybe we don't need both

JD: We already have this in the list

RS: I don't think we need the text Michael

MC: I think it's useful to have it in the text
... Still something you need to support
... Can be removed down the road

JD: Other specs have that in the conformance part

MC: I think it's useful to keep it in the conformance part
... Let's keep both

<joanie> +1 to keeping both

<ShaneM> +1 but I would have 3.5 point to the definition

<ShaneM> remember that the ID for a definition is dfn-<term>

<joanie> https://rawgit.com/w3c/aria/master/aria/aria.html#aria-grabbed links to the term

RS: Any objections?

<ShaneM> +1 to having the term point to the conformance requirement. It might also be possible to have respec notice if a link is broken.

PROPOSAL: Accept Michael's enhancement to the specification and add text to the term for definition to reference the normative text for…

<ShaneM> Rawgit is showing it now. You might also put <a>deprecated</a> in the conformance section - so they link both ways

AB: Shows up in every spec?

MC: Only those that link to it

PROPOSAL: Accept Michael's enhancement to the specification and add text to the term for definition to reference the normative text for deprecated.
... Accept Michael's enhancement to the specification and add text to the term for deprecated to reference the normative text for deprecated.

RESOLUTION: Accept Michael's enhancement to the specification and add text to the term for deprecated to reference the conformance subsection on deprecated.

Action 1723

action-1723

<trackbot> action-1723 -- Joanmarie Diggs to Editorial - create sections listing the roles that provide (1) nameFrom:author and (2) nameFrom:contents -- due 2016-02-24 -- OPEN

<trackbot> http://www.w3.org/WAI/ARIA/track/actions/1723

Action 1743

action-1743

<trackbot> action-1743 -- Joanmarie Diggs to Put aria-activedescendant on application and request wg review -- due 2016-02-17 -- PENDINGREVIEW

<trackbot> http://www.w3.org/WAI/ARIA/track/actions/1743

JD: That's done

RS: Addition of attr
... Any text added?

<fesch> https://github.com/w3c/aria/commit/f7220c213

*crickets*

<ShaneM> 3.2.97 works

*discussion about ReSpec breaking the web*

JD: Before under application
... It had a list of three different things you can do
... *reading diff out loud*
... We have to tell them what to do, and they have to manage focus
... Questions about that?

<joanie> Manage focus of descendants as described in Managing Focus, updating the value of aria-activedescendant to reference the element containing the focused content.

<Rich> https://rawgit.com/w3c/aria/ACTION-1730/aria/aria.html#deprecated

<ShaneM> old versions of ReSpec are NOT on the W3C Tools server. I just checked. UGGH

<MichaelC> Reachable from rawgit URI: https://raw.githubusercontent.com/w3c/respec/develop/builds/respec-w3c-common-3.2.97.js

<MichaelC> We could make a global change in our repository, and revert it when respec is fixed

RS: You want to have one active descendant within the container

<MichaelC> with lots of colorful language in the commit message of course

RS: You can have it on a listbox

JS: Where is the real focus?

RS: Would be on the combobox

JS: Does that have active descendant?

BG: Point to role=option
... You're supposed to point at role=option, otherwise it's an author error

RS: If we have active descendant on role=application, we could have a problem

BG: If you're in role=application, and you're in role=textbox, you'd have to script the whole thing in there. And it wouldn't work on mobile anyway
... Where as if you just set focus to it, you get that for free.

BG so, spec that out.

RS: I have concerns about that.

<jongund> I can

<jongund> JN: Active descendant doesn't apply

<scribe> scribe: jongund

BG: No way to identify the widget
... There is no reference to another types of widget, when role=application

JS: When it is a list box you know the role of the active descendant is an option

JN: I am not sure hwy that matters

BG: The SR only know what it is

JN: SR will read the role of what is being pointing to with aria-activedescendant

RS: Would take precedence over the other one
... If you have it on role application, it can point to anything

JN: What every has focus will be what active-descendent
... Nested activedescendant would be wierd

RS: yes

BG: Anywhere in the combo box

JN: Special cases for combobox is special case, we should be consistent

BG: If we allow for combobox they will start to use it other places
... Having active-descendant work different on combobox
... It will hijack ....

Joanie: I don't get the example, I thought we were talking about AD on application

BG: If you have role=combobox on ancestor, the input box has focus
... The browser can only pick one

JG: Can we change the definition of combox

BG: I think we are making it much more complicated than it needs to be

RS: It just happens to work on the input box

Joanie: What happens when you have multiple elements with AD
... What if you have a custom list box with AD

BG: If you press that native .... would have to be scripted in
... It you used role=application it would not need to go to the edit field

<janina> +1 to Amelia

ABR: I haven't done alot of building widget, if something has AD but not focus is it is ignored

JS: I believe that too
... Dojo widgets seem to use AD and it used AD to maintain focus state of complex widgets

Janina: I wanted that feature, AL said it was tough to do

Joanie: Depends on the widget

JS: When you reload a page, unless it sitting in the cache

Joanie: There is some caching
... What is the next step

Jania: I have to leave early

JS: Let's think about this for another week
... It would be good if matt were here

RS: This is the other action, ProSpec what strong semantics can be overridden

<Rich> https://www.w3.org/WAI/ARIA/track/actions/1489

MC: I was not sure the specific issue

RS: I don't think we should let the plateform should be able to override
... aria-label, aria-labelledby, may aria-describeby

MC: I am not sure why this is an action to me, is there anything else we want to protect?

JS: aria-checked (tri-state)

RS: We have a desc, in SVG
... Authors have to put these in

JN: They should implement the naming algorithms

Amelia: It is a matter of priorties
... aria-labelledby takes precedence... makes sure that there is a reference to the calculation spec

RS: I think we want consistent priorities for SVG

FE: ... mapping

JS: AXAPI mapping of a name property depends on whether the label is visible; for other AAPIs, it's the accessible name property period

RS: Let give you a new date

MC: Let's put next week at the date, but it is lower priority than pub stuff

RS: CS on password role, send a note to list
... MS doesn't seem to have any issue

Joanie: JF wants edits in the text, longer

RSL We will talk about next week, may not make the date

<clown> action-1513?

<trackbot> action-1513 -- Michael Cooper to Set up ARIA 1.1 Requirements draft -- due 2016-02-25 -- PENDINGREVIEW

<trackbot> http://www.w3.org/WAI/ARIA/track/actions/1513

MC: I think I am done

<MichaelC> http://w3c.github.io/aria/requirements/aria-requirements.html

MC: The requirements had existed, but I populated them with some content
... .... reading updates test ....
... DId I say anything we should not being doing

JS: There are 6 requirements

MC: The longdesc is removed from HTML5, but is in an extension

RS: I think you need to say something about SVG
... SVG2 is using ARIA 1.1 into their spec
... We are trying to help them support ARIA 1.1

Amelia: It needs to include SVG host language

<clown> "extended to SVG 2.0"?

MC: I am trying to find the branch for this, missed comments

RS: I am calling it a day

<clown> "extended to include SVG2.0"?

RS: SVG is part of what we are doing

MC: Can people repeat that

Amelia: I will draft text

RS: I am calling it a day

rrsgant, draft minutes

- DRAFT -

ARIA Working Group

03 Mar 2016

Attendees

Contents