Rw: [Standards] compliance levels for 2008

Mridul Muralidharan mridul at sun.com
Mon May 14 22:03:38 UTC 2007


Tomasz Sterna wrote:
> Dnia 14-05-2007, Pn o godzinie 12:17 -0700, JD Conley napisał(a):
>> Also, the component protocol needs to be updated to be more XMPP
>> friendly and secure (AKA use stream:features and SASL),
> 
> Agreed.
> jabberd2 component protocol is a good basis for this.
> 
> 
>>  as well as be
>> coupled with a suite of protocols for performing common tasks
>> component
>> writers crave like: GetAllBareJIDResourcePresences, [...]
> 
> Do we really need all these?
> 
> I don't see a reason for anything more than authenticated hooking into
> packet router and binding a set of domains to route into the component.
> Component does it's own packet handling.
> 
> Anything more is something like a component-framework put in a server.
> It's a JCR, jadperl etc. component job, not server's.
> 
> 

Component protocol has been extended in way too many ways for all sorts 
of purposes - including proprietary extensions for trusted functional 
enhancements : where a private contract exists between the server & the 
component.

Most of these evolved because of a lack of documented approach on how to 
go about doing these things for way too long ... and now there are too 
many implementations/deployments out there with its own custom 
implementations : each of which would be optimal from its own design 
point of view, but might not fit the world view of others.
So do we need a spec which proposes to standardize one or some 
implementations now ? It might not be very optimal and I am not sure of 
the adoption of such a spec.
But we might get something interesting out of it.

Regards,
Mridul



More information about the Standards mailing list