[Standards-JIG] Re: Roster Subscription Synchronisation

James Bunton james at delx.cjb.net
Sat Sep 11 12:35:29 UTC 2004

Richard Dobson wrote:
> Is it just me or is this protocol not virtually identical to just using 
> JEP-0093?.
> The only thing missing from JEP-0093 is the use case explaining how to use 
> JEP-0093 with transports. Whats the point in it being like this?
No it's identical. JEP-0093 can't actually do the Roster Subscription 
Synchronisation. All it can do is push out contacts, it says nothing about 
their subscription status on the legacy network.

> If you really want to have the information attached to the presence (still 
> the wrong place IMO) in a way that is virtually identical to JEP-0093 you 
> should at least use JEP-0093 for this instead of inventing your own new 
> protocol element.
I don't know what you mean here. Do you mean using the JEP0093 namespace as 
the namespace for the x tag? I don't think that makes much sense, because 
JEP0093 allows you to send multiple item tags, whereas this protocol 
specifically states that you mustn't. Only one item tag per presence packet.

The main difference between this protocol and JEP0093 is that -93 is designed 
to push contacts out to the user. This protocol is designed to be an 
extension to the existing presence mechanism, to let a client know what 
subscriptions have already been made on the legacy service by its user.

Tijl Houtbeckers put it very succinctly in an email to me:
> What the <import/> tag really tells you is: "this is a subscription  
> request packet you already approved on another network, thus there is no  
> point in denying it. even if you deny it, people on that network will  
> still see you as long as the transport sees your presence". A JEP-0093  
> packet says nothing.. and *should* say nothing about presence  
> subscriptions.

Hopefully that will clear up why JEP-0093 isn't a good fit for what I'm 
actually trying to do.



More information about the Standards mailing list