[Standards-JIG] RE: Standards-JIG] Jingle vs. Zoep

Jean-Louis Seguineau jean-louis.seguineau at laposte.net
Fri Feb 10 08:37:55 UTC 2006


Hey Peter, 

IMHO you lend yourself well to criticism here. I am just pointing out that
whoever has tried to bridge XMPP and SIP know perfectly well that the issues
do not lie with the mapping of XMPP stanzas and SIP headers/content. And
frankly speaking one can implement a perfectly functional gateway without
even implementing anything described in the draft. As long as what ever is
on the SIP side of the gateway looks like SIP to other SIP UA, and as long
as what is on the XMPP side looks like XMPP to other XMPP client, this is
all what is needed.

Trying to make anybody on this list believe that XMPP-to-SIP gatewaying is
only a matter of mapping would be misleading, isn't it? Looking at the
freely available gateways (yes there are such XMPP/SIP gateways), I would
not say that they have stormed the world and made XMPP/SIP integration
seamless.

A proper gateway would require much more than simply mapping JIDs to SIP
URIs. It would require much more than simply mapping PIDF to presence. I
believe I already explained elsewhere how these two protocol are trying to
solve different issues. Even the nice people at Zoep will soon find out...

Lastly, I am probably amongst the few people on that list that can pretend
having become acquainted with a little more than the 250 pages mentioned
elsewhere on the list (there is a little as circa 66 RFC about SIP/SIMPLE,
not to mention the drafts) as well as the application of them in other
standard bodies such as 3GPP. So excuse me for saying that 'you', the
authors, might benefit from this meager experience. And the community may
even gain more if the potential issues are exposed early. Now if you want to
keep Jingle a closed subject amongst yourself, there is nothing much we can
do about it, isn't it? 

Along the same lines, IMHO tying Jingle whit libjingle is equally
misleading. And it looks to me as not the proper way to design the protocol.
The library must be the result of the protocol, not the other way round. And
the fact that libjingle or that a service based on this technology exists is
by no mean a good enough reason to force the protocol in a particular
direction as this is already happening. (I have less experience on the low
level implications of the current protocol, but I'm sure others than just
the authors may want to share their own views on how Jingle may be improved)

Building a proper protocol is not easy. And everybody on this list is
entitled to bring different angles to its design. And I agree with other
list members that having no formal process similar to what IETF or other
standard bodies have, where workgroups start by laying out their missions
and plan openly how they would go about it, start to show the limit of
current way of bringing enhancements to XMPP.

Nobody here is saying anything other than there are ways for improvement
through a continuous process of open critical assessment... Critical
assessment has nothing personal about it, AFAIK.

Jean-Louis

-----Original Message-----
Message: 2
Date: Thu, 09 Feb 2006 18:53:07 -0700
From: Peter Saint-Andre <stpeter at jabber.org>
Subject: Re: [Standards-JIG] Jingle vs. Zoep
To: Jabber protocol discussion list <standards-jig at jabber.org>
Message-ID: <43EBF203.2000103 at jabber.org>
Content-Type: text/plain; charset="iso-8859-1"

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Jean-Louis Seguineau wrote:
> I don't feel that bridging XMPP/SIP through a gateway is that of an issue
> for the most common features of SIP (call or message sessions) and SIMPLE
> (presence) It requires a little more work than what is describe in Peter's
> draft ;)

If you have provided feedback on draft-saintandre-xmpp-simple and I have
neglected to incorporate it, please do let me know. :-) Naturally we'll
have to define something similar for signalling and media negotiation,
since draft-saintandre-xmpp-simple is just for basic messaging and
presence (and I do mean basic -- we talk only about "page-mode"
messaging on the SIMPLE side, not MSRP). Once we (the Jingle co-authors)
get the Jingle specs to a more stable state, we have every intention of
specifying how to gateway between the Jingle and SIP. At that point
feedback from those with heavy SIP experience will be much appreciated. :-)

Peter




More information about the Standards mailing list