[Standards-JIG] Dead participants in MU-conf, JEP-0045

Heiner Wolf wolf at bluehands.de
Fri Dec 10 17:40:02 UTC 2004


>> Godd if it doesm but not enought. There are clients which do not
answer
>> to my jabber:iq:version. No response, no error. The same would happen
to
>> MUC if MUC would ask. So relying on the error is not enough.
Moreover:
>> there is message-traffic in the room. The messages are also routed to
>> these dead resources. But they are not removed. So I guess they don't
>> generate errors. They probably perish like my jabber:iq:version.
>
>If no error is returned then that is the fault of the server
>router/delivery mechanism in question, not the MUC service.  If an XMPP
>server can't reach the destination recipient's server or the
destination
>recipient directly it MUST or SHOULD (depending on the circumstances)
>return an error to the sender. 

:-)
Yes, you are right, this is the fault of one or the other router. But
this is a distributed service of heterogeneous components. Some of them
fail, some may even do worse. MUC must not rely on the good behaviour of
all components. A chat might become unusable, because some servers are
badly implemented. Thats what I am experiencing. I have rooms, with
zombies. People avoid these rooms. I can not force my users to use only
correctly implemented routers/delivery mechanisms. For users its "the
Jabber chat" that has zombies. I am not happy with that, but it's the
real world. MUC should and could protect itself. 

hw
--
Dr. Klaus H. Wolf
bluehands GmbH & Co.mmunication KG
http://www.bluehands.de/people/hw
+49 (0721) 16108 75
--
Jabber enabled Virtual Presence on the Web: http://www.lluna.de/
Open Source Future History: http://www.galactic-developments.com/



More information about the Standards mailing list