RE: Requestless notifications?

update - we discussed on the call and agreed that this is not necessary for the speech protocol.



________________________________
From: Young, Milan [Milan.Young@nuance.com]
Sent: Wednesday, July 27, 2011 7:07 PM
To: Robert Brown
Cc: HTML Speech XG
Subject: RE: Requestless notifications?

The feature is useful in the context of MMI.  In this scenario, the browser would contain various modality component(s), and the service would perform the role of the interaction manager.  In such cases, the service might need to inform the modalities within the browser that an external modality has generated an event.

As suggested in my original text, I agree that this feature is not necessary to implement MMI.  The browser could use a separate WebSocket protocol, for example, to transmit/receive similar information.  But I figured since we already had a WebSocket open, why not re-use the channel?  It doesn’t seem to add any complications to the protocol.

I also object to the characterization that these notifications are “unsolicited”.  Like the service is sending the user credit card offers or something :).  In practice, these notifications would have been setup by parameter or query string on the URL.

Let’s discuss more on tomorrow’s call.

Thanks

________________________________
From: Robert Brown [mailto:Robert.Brown@microsoft.com]
Sent: Wednesday, July 27, 2011 4:39 PM
To: Young, Milan
Cc: HTML Speech XG
Subject: Requestless notifications?

Hi Milan, we discussed this while you were out:
http://lists.w3.org/Archives/Public/public-xg-htmlspeech/2011Jul/att-0025/speech-protocol-draft-03-r3.html#notifications

Is there a compelling reason to keep it? If not, we’d prefer to remove it.