Re: ISSUE-636 ACTION-1398 Provide spec. text for aria-roledescription

On Apr 3, 2014, at 12:10 PM, Joanmarie Diggs <jdiggs@igalia.com> wrote:

> On 04/03/2014 02:36 PM, James Craig wrote:
>> On Apr 3, 2014, at 11:07 AM, Joanmarie Diggs <jdiggs@igalia.com> wrote:
>> 
>>>  I say at least, because when you use a real role, the screen reader
>>>  can do things like: a) include the number of charts (or videos) as
>>>  part of its "document summary" feature. b) provide quick keys to
>>>  jump to the next/prev chart (or video). c) respect the user's braille
>>>  verbosity setting, which may include not displaying certain roles in
>>>  braille, or displaying them via a symbolic indicator, or displaying
>>>  them using a localized abbreviation. Bit of a shame to lose that....
> 
>> I object to the insinuation that we're losing something; we can't
>> lose something we've never had. No one is implying we should use role
>> description *instead of* adding new roles. No one is arguing against
>> the benefit of having real roles with additional functionality. We
>> are all in agreement that that many new roles should be added, for
>> all the reasons you listed and many more.
> 
> I feel like we have had a miscommunication, and I am really, really
> sorry for that. And doubly so because I most certainly did not mean to
> insinuate anything, including what you stated above.

Ah yes. We did have a miscommunication. Sorry for insinuating the insinuation. ;-)

> To be clear (I hope): I am not opposed to the inclusion of localized
> role descriptions.

That was unclear to me. Thank you for clarifying.

> The point I have been trying to make is that while
> localized role descriptions make sense for the reasons you have outlined
> -- and therefore we should do it for those reasons -- ATK happens to
> already have a way to accomplish that (and a bit more nice-to-haves
> along with it). Thus I would really like to find a way where:
> 
> 1. All platforms can have localized role descriptions, and
> 2. ATK can have a mapping to platform roles (like ATK_ROLE_VIDEO)
>   when reasonably possible.

Yes. I don't think this is limited to ATK. Most platform accessibility APIs have a superset of the roles provided by ARIA 1.0.

> If we can accomplish 2, then "we" (ATK/AT-SPI) don't lose the additional
> functionality that our (ATK/AT-SPI) platform makes possible for us
> (ATK/AT-SPI).
> 
> It was/is my hope that there is some way we can accomplish that
> platform-specific mapping **without in any way interfering with
> localized role descriptions**. That is what I have been asking about,

Thank you for clarifying and I agree. Perhaps we can limit usage (or recommend against usage) of @role and @aria-roledescription on elements where there is no one-to-one mapping to an ARIA role. <video> is currently one of these cases.

> and I even tossed out an idea for consideration which would use the
> tokenized roles in a mail I sent before seeing your reply above. I'm
> trying to brainstorm for a solution....

The tokenized role approach is already available in ARIA 1.0. In other words, you can do role="button link" and the computed role will be button. If and when the video roles is standardized, authors will be able to use role="video group" the way you've described.

> Again, my sincerest apologies for any offense I may have caused.

No offense taken. Mutual apologies for the misunderstanding. Welcome to the W3C mailing lists, where many things are misconstrued, but we usually end up with the right solution. ;-)

Cheers,
James

Received on Thursday, 3 April 2014 20:16:16 UTC