Re: [aapi] ARIAAttrModified?

Hi Becky,

The idea is that when exposing an accessible object for a given DOM
node, call it "foo", then if foo has one of these 'writeable' aria
attributes (http://www.w3.org/TR/wai-aria-practices/#aria-write) that
are considered handy for an AT to change, then the accessible object
exposed can also expose one or more accessible actions that correspond
to the values (probably NMTOKENS), that the aria attribute could have.
If an AT invoked one of the actions then FF would update the
aria-attribute with a value that corresponds to that action.

So in your example, focus wouldn't be moved around (ideally), there
would just be an attribute mutation listener that caused the grid to
sort appropriately. It requires a mind shift from thinking of the aria
attribute as passive semantics, but instead, more of an active player.
The execution path is such that the aria-attribute mutations drive the
behaviour. In some ways it might lead to cleaner web app code. but I
admit I'm skeptical about adoption of this design pattern. If I wasn't
I'd be convinced this is a great direction.

I'm not seeing black or white on this.

cheers,
David
Becky Gibson wrote:
> I guess I really don't understand the use case for this?  When/how would I 
> expect an AT or a browser to change an ARIA property?  I suspect this may 
> have to do with alternative input, such as voice, which I admittedly know 
> little about.   For example, I can't imagine coding the Dojo Grid to 
> listen for an aria-sort property change and then take the appropriate 
> action to re-render the grid.   Given voice input perhaps this would be:
> "focus the Country Column Header"
> "sort ascending"
>
> The following would happen in response to the above voice commands:   the 
> column header labeled "Country" would be focused; then the sort ascending 
> action would be performed and the grid re-rendered;  then focus would be 
> put back into the column header.  Is this an expected use case?   Can 
> someone provide another?  Right now I'm having a hard time accepting that 
> we need separate DOM Mutation events just for ARIA. 
>
> -becky
>
>
> Becky Gibson
> Web Accessibility Architect
>                                                        
> IBM Emerging Internet Technologies
> 5 Technology Park Drive
> Westford, MA 01886
> Voice: 978 399-6101; t/l 333-6101
> Email: gibsonb@us.ibm.com
> blog: weba11y.com/blog
>
>

Received on Friday, 6 March 2009 02:05:29 UTC