let's move on

[eostrom@acm.org (Erik Ostrom)]

let's finish this thing off.

there are two levels of change requests that have been made:
controversial and uncontroversial.  some uncontroversial requests have
been made but not incorporated into the spec.  those should be pretty
easy to take care of.

under "controversial", my impression is that we have two basic types
of requests:  "this isn't specified and it should be", and "this is
specified and i don't like the way it's specified".

could everyone who cares look over the spec and figure out:

  * are there things yet unspecified that really should be?
  * are there things you don't like that are deal-breakers?

we're not going to come up with a protocol that everyone thinks is
ideal; what i'd like is to have a protocol that we can all agree to
implement (and one that we can build on top of to do the more ideal
things that we'd like).  and to have it before the millenium.

so, is there stuff you really can't stomach in the spec?  are there
things you want to implement that you can't reasonably do on top of
the current spec?  or are we done?

--erik