Re: Managed states related to aria states (was "Re: [aapi] UAI TF Meeting Agenda Tue Feb 24 2015").

So it's not supposed to include states for things like HTML select?

On Tue, Mar 17, 2015 at 11:47 AM, David Bolter <dbolter@mozilla.com> wrote:

> Trying to get my head back into this context. My understanding of an
> example of browser managed state related to ARIA would be something like:
>
> https://dxr.mozilla.org/mozilla-central/source/accessible/generic/Accessible.cpp#1196
>
> I.e. adding the FOCUSABLE state to children of an aria-activedescendant
> container.
>
> Cheers,
> David
>
> On Tue, Mar 17, 2015 at 11:40 AM, Alexander Surkov <
> surkov.alexander@gmail.com> wrote:
>
>>
>>
>> On Tue, Mar 17, 2015 at 10:46 AM, Joseph Scheuhammer <clown@alum.mit.edu>
>> wrote:
>>
>>> On 2015-03-16 5:23 PM, Alexander Surkov wrote:
>>>
>>>> So, unmanaged is something that ARIA widget author should take care,
>>>> managed is something that the browser takes care of. Is that correct?
>>>>
>>>
>>> Close.  Two things:  (1) In this context, there are no "unmanaged"
>>> states, despite what the glossary entry says.   (2) The issue is not
>>> limited to ARIA.
>>>
>>> The distinction has to do with who or what is controlling the states.
>>> One case is where the author manages the states, for example, setting the
>>> @selected attribute on <option> elements within a <select> using script.
>>> The other case is where those states are managed by the user agent.  In
>>> either case, the relevant AAPI states are the same, regardless of what is
>>> managing them.  With respect to selection, that would be
>>> MSAA/IA2:STATE_SYSTEM_SELECTED, UIA:IsSelected, ATK:STATE_SELECTED, and
>>> AXAPI:AXSelected.
>>>
>>>
>> I thought the whole point of splitting states into these two categories
>> was that the states defined by ARIA are already mapped and you need to
>> complete the a11y mapping by describing what states comes from the browser.
>>
>>
>>>  How does this part conform to HTML a11y mapping guide? I assume you
>>>> don't want to list unmanaged states for, say, HTML select, no?
>>>>
>>>
>>> I don't understand the relevance of the HTML a11y mapping spec here.
>>> What section(s) of that spec bear on conformance with respect to this
>>> issue?  The task David took on was to create a list of AAPI states that are
>>> managed by the user agent.
>>
>>
>> HTML is mostly unique thing where the browser maintained states comes
>> from. If you don't want describe things in HTML terms then you need to have
>> sort of universal neutral language. Not sure how to approach to it.
>>
>>
>>>
>>>
>>> --
>>> ;;;;joseph.
>>>
>>> 'Array(16).join("wat" - 1) + " Batman!"'
>>>            - G. Bernhardt -
>>>
>>>
>>
>