Re: hit testing and retained graphics

On 7/2/2011 6:54 PM, Benjamin Hawkes-Lewis wrote:
> On Sat, Jul 2, 2011 at 3:06 PM, Steve Faulkner
> <faulkner.steve@gmail.com>  wrote:
>> I don't see why you would have to duplicate the whole accessibility
>> stack  to provide focus tracking for a screen magnifier, can you
>> explain this a bit further?
> The remote system access server would need to translate the remote
> applications (as accessed by the accessibility tree plus custom hooks)
> into DOM. To support custom views/controls for which we do not have
> semantics in the web stack or to provide any application-specific
> customizations, local AT would have to make special interpretations of
> the DOM (either directly or as exposed to the accessibility API). Thus,
> the accessibility stack (converting remote applications into accessible
> interfaces) would need to be duplicated.
>
> If you disagree, can you explain precisely what you think the remote
> system access server on the one hand, and local AT on the other, would
> need to do?
They can limit their protocol to sending information to elements which 
gain focus,
for high-latency/bandwidth constrained environments; they could enable
a protocol along websockets which allows them to pass platform specific
commands for Accessibility tree queries (such as Microsoft's UIA or 
Apple's UIAutomation);
they could do an arbitrary amount of introspection and heuristics.

It's the client side application, the one in ECMAScript, which does the 
DOM manipulation.
And it's the user agent that has to get the ARIA / Accessibility tree 
that's provided by the DOM
working in a manner that ATs can use.

Received on Sunday, 3 July 2011 02:34:02 UTC