RE: voiceURI usage

Hello Eitan,

The URIs in the constructors to speech recognition and synthesis were intended for both:
  * Internal browser consumption (the opaque string you mention)
  * A means of pointing the host browser to an external speech service.  The browser would need some published protocol to communicate with the remote.

I request we preserve both use cases.

Regards


> -----Original Message-----
> From: Eitan Isaacson [mailto:eitan@mozilla.com]
> Sent: Wednesday, January 23, 2013 11:24 AM
> To: public-speech-api@w3.org
> Subject: voiceURI usage
> 
> Hi.
> 
> I have been trying to wrap my head around the use of voiceURI, both as an
> attribute of SpeechSynthesisVoice and SpeechSynthesisUtterance.
> 
> It seems to me that a UA implementation would have some internal (and
> potentially extendable) speech services. So as I understand it, the URI would
> only have internal meaning to the UA, and should really be treated as an
> opaque immutable token by content scripts. If this is the case, the usage of
> voiceURI is unclear to me, and the suggestion of removing voiceURI as filed in
> bug 20529[1] makes sense.
> 
> On the other hand, maybe I don't understand how voiceURIs should be used.
> For example, could you optionally provide the speech service part of the uri
> without the trailing voice identifier and have the service choose the default
> voice for that service (this differs from the global UA default voice)? Or could
> you provide an HTTP URI in a SpeechSynthesisUtterance where some standard
> GET request would be used to retrieve the synthesized audio?
> 
> Could someone please provide examples of what voiceURIs might look like?
> Both for local and remote services.
> 
> Cheers,
>   Eitan.
> 
> 1. https://www.w3.org/Bugs/Public/show_bug.cgi?id=20529
>

Received on Thursday, 24 January 2013 23:07:52 UTC