Re: Use of BSTR in MSAA VARIANT

Hi, Jason.

Neither MSAA nor IA2 suggest better than BSTR hack to expose semantics of
those elements. That means either generic roles work for those elements
just fine or there's no much well-known examples where generic role fails.
If the latter case is true then the way IA2 evolves will be something
different than BSTR hack I think. So I think I agree there's no need to
make the spec saying to use BSTR hack. However we have two implementations
of BSTR approach, I'm not aware of other implementations that don't make
it. So if the spec says something that nobody follows then it doens't
really make sense. Also this kind of forking will complicate the spec.
Thanks.
ALe.x


On Tue, Mar 4, 2014 at 12:13 AM, jason@accessibleculture.org <
jason@accessibleculture.org> wrote:

> Hi Alexander,
>
> On 4/03/2014, at 10:50 am, Alexander Surkov <surkov.alexander@gmail.com>
> wrote:
>
> Hi. I don't see any practical benefits of dropping BSTR hack. Also I would
> avoid to wake up a sleeping dog, I'm not sure who may have dependencies on
> it.
>
>
> I guess the question for me is more whether or not the mapping guide
> should be specifying and recommending a hack that is implemented by one or
> two browsers (are there others that use this hack?) and discouraged by the
> keepers of the MSAA spec. I've understood the guide to be recommending
> mappings first, and documenting what certain browsers do second, and only
> to the degree that what these certain browsers do aligns with or informs
> the recommended mappings.
>
> Also, even if the mapping guide dropped the hack, that wouldn't force
> Firefox to do so.
>
> Cheers,
>
> Jason
>
>
> Thanks.
> Alexander.
>
>
> On Sun, Mar 2, 2014 at 4:18 PM, jason@accessibleculture.org <
> jason@accessibleculture.org> wrote:
>
>> Looking for folks' opinions on how we relate, in the HTML to
>> accessibility API mappings, Firefox and Chrome's use of VARIANT to return
>> the tag name as string (BSTR) for elements without established roles in
>> MSAA.
>>
>> I think we had decided at one point that since this approach isn't
>> "described by the MSAA specification", we wouldn't indicate or promote it
>> as a preferred mapping in the actual mapping tables, but instead just
>> include a note about it [1]. The note exists [2], but we still have this
>> use of BSTR noted in the MSAA + IA2 mappings themselves, for example, see
>> the mapping for abbr [3].
>>
>> Am I right in thinking, one, that Mozilla is looking at dropping the
>> "BSTR hack" [4], and two, that the individual element mappings for MSAA +
>> IA2 shouldn't include this use of BSTR?
>>
>> Cheers,
>>
>> Jason
>>
>> [1] https://www.w3.org/Bugs/Public/show_bug.cgi?id=16769#c3
>> [2]
>> http://rawgithub.com/w3c/html-api-map/master/index.html#use-of-msaa-variant-by-some-user-agents
>> [3] http://rawgithub.com/w3c/html-api-map/master/index.html#el-abbr
>> [4] https://bugzilla.mozilla.org/show_bug.cgi?id=798492
>>
>> Jason Kiss
>> jason@accessibleculture.org
>> http://www.accessibleculture.org
>>
>>
>>
>>
>>
>
>

Received on Tuesday, 4 March 2014 16:53:14 UTC