Marc suggested:
I any case, may I suggest a Bar BOF in Prague? Plotting revolutions in
coffeehouses is a very old tradition.
Excellent idea. Perhaps this should be plotted over jasmine tea instead of
coffee...
The point I really want to stress is that we must stop deliberately
designing privacy, integrity, and obscurity weakness into our protocols,
and where we can't avoid weakness we should at least consider its
implications. We have a real lack of understanding of these issues in the
community. For example, if Alice and Bob have a communications session, IETF
has never clued onto the fact that Alice and Bob might want intermediary
Charlie not jut to be unable to read the data of their session, but to not
even be able to know that they have one. We might not be able to hide the
fact that Alice has a session with SOMEBODY from her next-door neighbor
Allen, or the fact that Bob has a session from his next-door neighbor Burt,
but even if Allen and Burt are working together, we should be able to hide
the Alice-Bob relationship.
What do I mean by not designing weakness into our protocols? I give you SIP,
for example. After twelve years of work, I have yet to make a real call
using the optional "sips" signaling model. Why? It's optional. Nobody uses
it. Actually, I'm having a hard time using even basic SIP any more -- it
looks like Google just pulled-the-plug on my telephony ISP service, which
had been provided by the Gizmo Project. But that's another problem.
--
Dean
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf