[jdev] The future of Jabber/XMPP?

[Dave Cridland]

On Mon Jul 12 19:43:00 2010, Bill de hÓra wrote:
> On Mon, 2010-07-12 at 10:03 +0100, Dave Cridland wrote:
> 
> > 5) In terms of Google specifically - Google is a large,  
> broad-based,
> > company with a momentum all of its own. Very much like Microsoft,
> > it's important to remain objective when looking at what they're
> > doing. So while Google have insisted (on multiple occasions) that
> > XMPP, using XML, is way too verbose (and therefore power hungry)  
> for
> > mobile, I'd note that by contrast Nokia's use of XMPP to the  
> handset
> > appears to be entirely standards-based.
> 
> It's not that XML is power hungry (streaming parses do ok cpu wise),
> it's that XMPP/XML eats bandwidth and the chatiness (no pun  
> intended) of
> XMPP when people aren't saying anything will tend to keep to impact  
> the
> phone radio (and thus the battery). Anyone tracing XMPP knows that  
> there
> is a lot of presence stuff flying about when people aren't saying
> anything and that part for mobile at least, is pretty inefficient.

Ah, yes. Google have an undocumented extension for buffering up  
presence until there's something else to send. I, too, have been  
thinking along these lines. It's vastly simpler to do (in both client  
and server) than SIFT. I've done a bit of looking into this, and it  
appears to save battery substantially.

> A well defined mobile XMPP profile seems like a good idea, instead  
> of
> grab bagging various XEPs and trying them out. Once there was an  
> optimal
> mobile profile, whether or not XML makes sense would be much  
> clearer.

 From a mobile client implementor's perspective yes.

 From a protocol designer's perspective, it's horrible - a Big Switch  
is very hard to specify, and often ends up including the kitchen sink.

This in turn means it's a nightmare for server implementors.

I'd be interested in seeing a mobile XMPP specification which - like  
the Lemonade one for email - simply points to a handful of  
specifications and how to use them effectively, as well as providing  
useful background. I have some of this written already, so I'll  
submit it as-is, and work on improving it.

Dave.
-- 
Dave Cridland - mailto:dave at cridland.net - xmpp:dwd at dave.cridland.net
  - acap://acap.dave.cridland.net/byowner/user/dwd/bookmarks/
  - http://dave.cridland.net/
Infotrope Polymer - ACAP, IMAP, ESMTP, and Lemonade