[standards-jig] gateway handling of legacy contact lists

maqi at jabberstudio.org maqi at jabberstudio.org
Sat Dec 13 11:36:44 UTC 2003

On Sat, 13 Dec 2003, Sebastiaan Deckers wrote:

> I agree with that statement.  Transports should mirror the remote roster
> -- not import them selectively.
> The usage of jabber:x:roster is going to make it more confusing and
> difficult for people to register with transports.  (And to some degree
> more difficult for client developers.)

At last! Someone else looking at the problem from a user's point of view
:-). Thanks, I thought I was the only one ;-).

I understand why protocol people don't want to embrace the type=subscribed
hack. But I don't understand why they say "jabber:x:roster is optimal" as
it's obviously NOT.

However, what about saying "jabber:x:roster has to suffice right now but
there is work to be done there"?

Some time ago (I'm too lazy now to look it up), there has been some
discussion on enhanced roster manipulation facilities. This includes
pushing and pulling the entire roster or parts of the roster by services
on the Jabber network. This would make a LOT of sense not only with
transports (which not only can use the pushing but also the pulling part
as for now, transports have to go through major pain to get a consistent
list of contacts they are expected to handle) but also for other services
(think of the often-mentioned business case: With pushing, you could
register with a "Company roster" service which updates your roster
accordingly all the time, transparently, without annoying jabber:x:roster
messages or "You have been unsubscribed" notices).


More information about the Standards mailing list