Re: telepathy roadmap and media streaming

On Thu, 28 Sep 2006, Robert McQueen wrote:
> I think abstracting stuff into seperate concepts of a video call and a
> voice call would be unwise, because at least in Jingle calls, individual
> streams can come and go during the lifetime of a call. Ie, you can be in
> a voice call and plug in a webcam and then offer a video stream to
> somebody. They could then go "oh yes what a good idea", and then plug
> their webcam in, and upgrade the video stream to bidirectional...

Doesn't this imply automatic video stream connection?  What if I connect a 
digital video recorder with a firewire cable.  This dvr happens to be 
supported as a webcam.  This doesn't mean I want to create a video stream 
with the person I'm currently talking.

The API I proposed would require the user to explicitly publish the video 
stream for the remote end to subscribe.  Granted, this may not be how all 
IM protocols work.  But I don't think it prevents the user from 
configuring gossip to automatically publish and accept subscriptions of 
the video stream.  So the API supports your use case scenerio.

> There's a reason that the Telepathy API exposes the streams to you, each
> with their own directionality. :)

My question is; should this API be exposed to the end user?  Or should it 
be hidden behind a user friendly interface that does most things 
automatically.

My goal is to develop an interface that allows the end user to make a vv 
transaction independent of the IM protocol being used.  I.e. the use case 
for this vv transaction should be the same for SIP, Yahoo Messenger and 
Jingle.  I'm not sure where the directionality of the stream would be 
useful to the end user.

--jc
-- 
Jimen Ching (WH6BRR)      [email protected]     [email protected]