Thanks for reminding me of that Jamie. What do you think of using
IAAction::doAction with IA2 specified negative numbers, e.g.
enum IA2ActionConstants {
IA2_ACTION_OPEN = -1, // Used to inform the server that the client
will signal via IA2_ACTION_COMPLETE when it has consumed the content
provided by the object. This action allows the object's server to wait
for all clients to signal their readiness for additional content. Any
form of content generation that requires synchronization with an AT
would require use of this action. One example is the generation of text
describing visual content not obvious from a video's sound track. In
this scenario the TTS or Braille output may take more time than the
related length of silence in the video's sound track.
IA2_ACTION_COMPLETE = -2, // Used by the client to inform the server
that it has consumed the most recent content provided by this object.
IA2_ACTION_CLOSE = -3 // Used to inform the server that the client no
longer requires synchronization.
}
Alternatively we can add a new method doNamedAction and use a BSTR input
parameter, but the above has no impact on the existing interface.
Pete
On 7/21/2011 7:53 PM, James Teh wrote:
> On 22/07/2011 10:31 AM, Pete Brunet wrote:
>> 2) IAccessibleAction and ARIA live regions should be used to synchronize
>> the streaming of text descriptions as follows:
> If IAccessibleAction is to be used like this, we probably need to
> optimise the interface to allow for action constants, similar to the
> proposal for relations. Having to iterate through all actions is fairly
> inefficient.
>
>> - The text cue is presented via a live region.
> There probably needs to be an additional attribute (or an additional
> value of the live attribute) to indicate that "open" and "resume"
> actions are required. Executing them otherwise is wasteful. Perhaps
> "live:interactive"?
>
>> - A "resume" action indicates completion of the presentation of the text
>> cue.
> Perhaps this should be generalised to just "presentationDone",
> "liveDone" or similar. "Resume" makes it very specific to
> pausing/unpausing. However, it seems to me that it should be generalised
> to allow other actions when the AT has finished presenting the live
> update to the user.
>
>> - An "open" action provides ref counting so that the video will not
>> resume playback until all users of the interface signal readiness to
>> proceed.
> To clarify, I assume both of these actions would be on the live region
> object? This makes the most sense to me.
>
> Jamie
>
_______________________________________________
Accessibility-ia2 mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/accessibility-ia2
