Re: RFC2109 addition...

Some of the recent discussion on cookie restrictions has
been wildly speculative. We risk writing a spec that puts
unrealistic constraints on implementations.

I'd like to ask people who are making proposals to qualify
their proposals with actual implementation experience, or
even some sample pseudo-code for exactly how this is supposed
to work in a client. The kinds of comments that have
been made recently about how we're just providing challenges
for implementors are really worrisome.

As a working group, we're spending WAY too much energy
on cookies, when there are an enormous number of other
issues that we *MUST* focus on.

It may be that for us to make progress we will need to spin
"state management" into a separate working group, so that
HTTP-WG can focus on the core HTTP protocol.

I've been resisting a push from the area directors that
we should shut down HTTP-WG and create new working groups
to deal with these sub-issues, but I could use some help
from you (HTTP-WG members) to keep focus.
--
http://www.parc.xerox.com/masinter