There are cases where the implementer is bridging from a toolkit and has
no control over that toolkit's structuring and IA2_RELATION_MEMBER_OF is
needed.  But do we need the reverse?  What is the use case for
IA2_RELATION_GROUPING_OBJECT_FOR?  Grouping objects don't get focus.  In
a virtual cursor scenario I haven't seen AT announce a group's members
when the cursor lands on the grouping object (but the converse relation
is needed when the cursor is on a contained accessible).

I also notice that noone has argued for adding
IA2_RELATION_POPUP_INITIATOR_FOR (the converse of
IA2_RELATION_POPUP_FOR).  I'll not include it if there is no
justification for it.

Pete

On 2/25/12 11:53 AM, Alexander Surkov wrote:
> I'd say if having a relation means either we don't have good
> assumptions of the hierarchy or it's more performant way than tree
> traversal. In either case we want to have opposite relation expect the
> case when related accessible is a child (not performant to run through
> children but performant to get a parent).
>
> IA2 says about member_of relation
> (http://accessibility.linuxfoundation.org/a11yspecs/ia2/docs/html/group__grp_relations.html#gb0213d9d1193fb6e6f50e1f2115aa911):
> It is also possible that each member has multiple additional targets,
> e.g. one for every other member in the group. That makes me think you
> can't do a good guess about hierarchies. And either we need to have
> opposite relation or fix a spec.
>
> Btw, Firefox exposes member_of on child accessibles from subtree of
> ARIA atomic regions only. I don't think opposite relation is
> applicable here. I think I agree with Peter Korn that we shouldn't
> expose relations in every case so if AT has a good idea about
> hierarchies then relations should be used as fallback when hierarchies
> fail. That's what we do for node_child_of relation for example.
>
> Thanks.
> Alex.
>
>
> On Sat, Feb 25, 2012 at 8:06 AM, James Teh <[email protected]> wrote:
>> On 25/02/2012 1:28 AM, Peter Korn wrote:
>>> It has been a LONG time since I wrote any screen reader code on Windows,
>>> but I recall clearly that the old Win32 radio button group box control
>>> was always a peer of the buttons it was grouping, not the parent.
>> Correct. However, if you're using something that implements IA2, someone has
>> obviously already gone to the trouble of implementing their own
>> accessibility objects, so they may as well sanitise the hierarchy as well
>> (i.e. make the grouping a parent). If anyone can come up with use cases
>> contradicting this, there's a good case for the new relation. Otherwise,
>> probably not.
>>
>> Jamie
>>
>> --
>> James Teh
>> Director, NV Access Limited
>> Email: [email protected]
>> Web site: http://www.nvaccess.org/
>> Phone: +61 7 5667 8372
>>
>> _______________________________________________
>> Accessibility-ia2 mailing list
>> [email protected]
>> https://lists.linuxfoundation.org/mailman/listinfo/accessibility-ia2
> _______________________________________________
> Accessibility-ia2 mailing list
> [email protected]
> https://lists.linuxfoundation.org/mailman/listinfo/accessibility-ia2
>

-- 
*Pete Brunet*
                                                                
a11ysoft - Accessibility Architecture and Development
(512) 467-4706 (work), (512) 689-4155 (cell)
Skype: pete.brunet
IM: ptbrunet (AOL, Google), [email protected] (MSN)
http://www.a11ysoft.com/about/
Ionosphere: WS4G
_______________________________________________
Accessibility-ia2 mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/accessibility-ia2

Reply via email to