[standards-jig] JEP-115 Redundancy?

Bart van Bragt jabber at vanbragt.com
Tue Sep 30 13:59:32 UTC 2003


Jacek Konieczny wrote:
> I think one of the goals of JEP-115 is to avoid queries which are known to
> give the same result. But the way it is defined in JEP-115 is not very
> secure.

A warning in advance: this is all coming from a pretty naive point of 
view. Haven't actually done anything with Disco etc :D

Anyway. I was wondering why the client is handling almost everything by 
itself in the current JEP. One of the really nice things in XMPP/Jabber 
is the 'simple client, complex server' philosophy. Almost all extensions 
that I'm seeing at the moment are heavily client based. I can see the 
point of that, it's much easier to upgrade a client than it is to 
upgrade a server but still...

Anyway. Wouldn't it be much more efficient if the server would retrieve 
and cache the capability information? IMO the server is in a _much_ 
better position to do this. It has to talk to a lot of clients anyway, 
it doesn't have the same bandwidth restrictions as clients have (mobile 
users etc) and it can cache this information for _all_ it's users 
instead of for one client that might even not be able to cache anything 
at all (Java applet, mobile client, etc).

Just a though :D

Bart





More information about the Standards mailing list