[Standards-JIG] Re: Summary of roster proposal points

Peter Millard pgmillard at gmail.com
Wed Sep 8 20:39:46 UTC 2004


On Wed, 8 Sep 2004 10:06:33 +0100, Richard Dobson <richard at dobson-i.net> wrote:
> Well if a client ignores the JEP-0093 stanza they will simply get prompted
> with all the subscription requests, the best way to implement this would be
> to send out the JEP-0093 packet before any subscriptions, then have a delay
> of say 30 seconds before you actually send all the pending subscriptions or
> start sending all of them out when you receive your last subscription
> request from the client indicating that they have finished adding the
> contacts that were in the JEP-0093 packet.

I must be missing something in this whole discussion. It seems like
we've already solved this problem. For those not following all of our
JEP's at home.. here is how this SHOULD work (IMO as a council
member):

1) Transport disco's the user's client (the transport should have
received presence from the user in question so it can do this).

2A) If that client supports -93, the transport sends a JEP-93 packet
containing all of the users.  Goto (3).

2B) If the client does not support -93, send a message to the user
telling them to get a better client :) Honestly, -93 is a big feature
that many client authors often over-look and it's darn simple to
implement. This could be the use-case that forces client authors to
imlpement this jep more widely.

3) Use Case Ends.

This doesn't take into account synchronizing jabber & remote-IM
rosters, but are we trying to solve that problem, or just the initial
import problem?

pgm.



More information about the Standards mailing list