Re: [rtcweb] Proposed message to send to the IETF rtcweb and W3C WebRTC working groups.

[Philipp Hancke <[e-mail]>]

(removed stox)

Am 22.07.2013 17:14, schrieb Iñaki Baz Castillo:
> Great. First thing you should complain about is the fact that current
> WebRTC specification makes unfeasible for a browser to use SDP-XML as
> defined by XEP-0167.

I said this before, but since you insist on repeating your argument i'll 
repeat mine: I have running code doing exactly that.

It's hard work and there are some points where this is PITA, I have 
discovered numerous bugs in chrome (and the jingle spec) along the way, 
but it's certainly not unfeasible.


So far, this mapping works using the elements already defined in XEPs 
0167, 0293, 0294 and 0320 even. 
http://tools.ietf.org/html/draft-ietf-rtcweb-jsep-03#section-4.1 has a 
quite concrete list of features of what functionality the mapping 
between SDP and the xmlish SD has to define.

I'd note that I consider jingle a better way to talk about the session 
description because it has a very clear separation between codec 
negotation and transport which has led to concepts like trickle-ice. 
Also, it defined actions like content-add which seem to have influenced 
unified plan.


I have yet to see the current API fail completly and have been using it 
in ways that were certainly not imagined, e.g. early transport warmup 
using PRANSWER (and I still owe feedback on that to the JSEP authors; 
ironically, the solution is more SDP in the API).