On Tue, Oct 12, 2010 at 11:36 AM, Iñaki Baz Castillo <i...@aliax.net> wrote:
> 2010/10/12 M. Ranganathan <mra...@gmail.com>:
>
>>  But let me ask
>>> you a question: Do you know XMPP protocol? There are hundreds of
>>> *good* XMPP implementations "available in internet" licensed as free
>>> software, and most of them interoperate very well with each other.
>>
>>
>> XMPP is much simpler than SIP and it does much less. When you start
>> putting in codec renegotiation routing and call transfers and such
>> into the picture, it starts to look a lot like an XML version of SIP.
>> SIP is complicated because it can do a LOT.
>
>
> Sure, I know it. But now take a look to presence specification in both
> XMPP and SIP.
> XMPP presence works.
> SIP presence is the *worst* design in the world (painful XCAP and XML
> document management???). And even worse, nobody knows what exactly to
> implement as IETF did an uncomplete set of specifications.
>
>
>
>>> So, perhaps "internet people" coding XMPP stuff are better than
>>> "internet people" coding SIP stuff? Or *perhaps* SIP is more much
>>> difficult and complex to implement than XMPP? Or perhaps XMPP is a toy
>>> and doesn't scale well as SIP does? (if so we should tell Google and
>>> Facebook that they must change their IM protocol ASAP).
>>
>>
>> XMPP does less with the screwy scenarios.  There, as yet no XMPP
>> business phone and there is a good reason for that.
>
> Because XMPP doesn't implement voice yet (well, it does but it's not
> widely implemented).
>
>
>> Perhaps there will soon be one, given google and facebook is in the
>> game. Then XMPP will start resembling SIP ( albeit in XML).
>
> SIP is good for voice. XMPP is good for presence /IM.
>
> When XMPP comes to voice it becomes complex (XML version of SIP). I agree.
> Then when SIP comes to presence it should be easy and robust as XMPP,
> right? For sure this is not true --> bad design, the worst and more
> stupid design in the world: SIMPLE/XCAP.
>
>
>
>> Then you can complain about how complex XMPP is.....
>
> I think you understood wrong my previous mail. Let me show an example:
>
> - In XMPP if a device publishes "available" presence status it means
> that it is reachable (in SIP terminology we'd say "it is registered").
> So it's impossible to publish "available" without being reachable, the
> server imposes this restriction.
>
> - In SIP this is not true. An UA can publish "available" while it is
> not registered (not reachable). Of course I mean "available" from SIP
> point of view, I don't care the stupid IETF vision of "available on
> mail".
> SIP allows very exotic escenarios as external presence user agents
> publishing information in behalf of the users, but when coming to 99%
> of cases (a simple user's device that makes/receives calls and
> publishes presence) SIP protocol gets too much complex.
>


 SIP/SIMPLE is over engineered for the task of Instant Messaging. I
have not seen any major deployments of SIP/SIMPLE.

The only reason why one may want to use SIP/SIMPLE in preference to
XMPP would be if one wanted to start an IM session in a SIP session
that one has opened for some other purpose such as a voice/video call.

XMPP makes much more sense for IM.

XMPP and SIP are fairly easy to combine using third party call control.



-- 
M. Ranganathan

_______________________________________________
Sip-implementors mailing list
Sip-implementors@lists.cs.columbia.edu
https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors

Reply via email to