[Standards-JIG] Re: proto-JEP: Metacontact Storage

Yann Le Boulanger asterix at lagaule.org
Fri Feb 24 16:35:05 UTC 2006


Remko Troncon wrote:
>>Could you describe this in more detail? It's not clear to me (1) what
>>the 'account' attribute identifies and (2) what account merging is.
> 
> 
> 'Account merging' means that you show all contacts of all accounts in one
> flat list (like Gaim does, unlike what Psi does).
> 
> Say you have an account on jabber.org and one on gmail.com. If you have a
> user 'stpeter at jabber.org' on one server, and 'stpeter at gmail.com' on the other,
> you want to put them in one metacontact. This metacontact crosses the account
> boundary (the contacts are on different accounts). So, the only way to
> store your metacontacts would be to choose one 'master' account which contains 
> the metacontact data, and with that data the annotation of what other account
> the contact belongs to. It's not clean IMO (what if there are 2 masters,
> how to choose the master, ...), but i think there's no other way out.
> 
> My question is: why does a metacontact need a JID ? Shouldn't it just
> group JIDs together in an ordered list ?
> 
> cheers,
> Remko
> 

I take again your exemple. in account u at jabber.org you have as contact 
stpeter at jabber.org, and in account u at gmail.com, you have as contact 
stpeter at gmail.com. Now you merge account. so your roster will show:

all accounts
|-stpeter at jabber.org
|-stpeter at gmail.com

now you want to set stpeter at gmail.com as a child contact of 
stpeter at jabber.org. you'll send on jabber.org server (where parent 
contact is):

<parent jid='stpeter at jabber.org'>
         <child jid='stpeter at gmail.com' account='u at gmail.com'/>
</parent>

The way I thought things when I wrote (and implemented that under Gajim) 
is that there is one parent which has one or more child(ren). I indeed 
don't speek about ordering child, cause I think all clients can do that 
the way it want: by status, by name, by what_they_want.

Yann



More information about the Standards mailing list