Re: a positive "no thanks" to cookies?

Ben Laurie writes:
 > 
 > Isn't this rather self-defeating? This mechanism could then be used
 > instead of the cookie.
 > 
 > Cheers,
 > 
 > Ben.
 > 

Yes, of course you're right, if it were per-cookie instead of per-server.
If there were a generic "don't send cookies" command, that would not convey
any state information.

As Marc Hedlund pointed out, we don't need anything in the protocol
for this at all -- the same UI mechanisms that would have to be built
for that could entirely handle the problem. 

Sorry to waste time by bringing this up -- the dangers of "thinking out loud"
in a group like this.

--Shel