Re: [rtcweb] Does ROAP mandate the on-the-wire format?

Let's say the RTCWeb API passed JSON objects like the ones in ROAP in and out of the PeerConnection object. (I will be arguing that is one thing we should consider).  At that oping you could write the SIP over webesockets in JS in the browsers. You might even find some useful info on how to do the mapping in draft-jennings-rtcweb-signaling-gateway-01

I think there has been a lot of talking past each other on this. In some cases ROAP over webesockets might be a protocol used to speak directly to a ROAP to SIP signaling GW. So I view ROAP over a well defined transport to be a on the wire protocol but certainly not the only on the wire protocol. Just one that some systems could use. 

On the other hand, if one does SIP in JS (or the browser), that works too. 

Hope that helps clarify. 

On Oct 21, 2011, at 10:43 , Iñaki Baz Castillo wrote:

> 2011/10/21 Hadriel Kaplan <HKaplan@acmepacket.com>:
>> Well since the Browser will never know whether ROAP is really to another Browser vs. the Javascript, then yes it could be used that way. (it will probably be painful, but possible)
> 
> The question comes because my wire signaling implementation (SIP over
> WebSocket) is that: SIP. So the JS sends and receives pure SIP
> messages (with SDP) over the wire.
> 
> If ROAP includes a real SDP (as the draft currently states) then I'm
> done as I just need to create a ROAP object (in JavaScript) and pass
> it the received SDP string (and also set some session status
> variables, which can be easily mapped from the received SIP
> request/response).
> 
> Hope I'm right and I don't have to drop the work of long months.
> 
> Regards.
> 
> -- 
> Iñaki Baz Castillo
> <ibc@aliax.net>
> _______________________________________________
> rtcweb mailing list
> rtcweb@ietf.org
> https://www.ietf.org/mailman/listinfo/rtcweb