Hi,
On Oct 5, 2008, at 8:35 PM, Pavel Simerda wrote:
But as it seems now... the new RFC breaks the current behavior that
was
not the best, but acceptable.
The new RFC recommends a new behavior, IMHO, the correct one.
It still allows for the old, IMHO broken, behavior.
I don't mind that the new RFC switches language so that a user
decision to use a specific resource is respected, I just think that
the use of resources for the purposes described so far sub-optimal,
and we would be better with other solutions.
The RFC should move us to better solutions than those in the past. Of
course, what is the best solution is what seems to be the current
discussion. I believe that the solution should be a session-resume
protocol with link level acks, a bit in the spirit of XEP-0198.
The use of resources to kick an old connection solves the multiple
connection problem, but not the reliability problem. Messages are
still lost.
Best regards,
And the new behavior is totally broken unless other (for now
optional, specified in XEP-198) means are used for disconnection
detection and stanza acks.
Sure, if your client doesn't support XEP-0198 or something like that,
the same-resource-kick is better than nothing, and I don't mind that
RFC392[01]bis servers support it as a legacy thing.
I just don't think we should recommend it anymore. It was the best we
could do then, but it is not a complete solution, and while 198 is
till young and untested, at least it tries to solve the whole problem.
That's pretty bad and will lead to a loss of trust that XMPP services
work at least reasonably reliably.
Actually, the reliability right now is already pretty bad on
unreliable networks, and the half-solutions of same-resource don't
improve that much.
We need a proper solution and the current best is 198.
Best regards,
--
Pedro Melo
Blog: http://www.simplicidade.org/notes/
XMPP ID: [EMAIL PROTECTED]
Use XMPP!
