Plans

[Scott_Fahlman@SEF1.SLISP.CS.CMU.EDU]

I have reserved Wean 1302 (same room as last term) from 4:30 to 5:30 on
Mondays for the coming term.  Actually, I have grabbed the 4:00 to 4:30
slot as well.  If there are no conflicts this term, maybe we can start the
meetings at four and get home a bit earlier.  The first meeting of 1992
will be Monday, Jan 13, at 4:30.

Barring any last minute snafus, Paul Gleichauf will be joining the group in
the Bill Chiles Memorial slot.

There's a strategic issue I'd like to discuss on Monday, so all of you
should start thinking about it now.  Hemlock is really nice in many ways,
but I think we would not have created it if Gnu Emacs were already
available.  It is my impression that only a half-dozen people outside the
group use Hemlock as their default editor -- maybe not even that many.  The
rest of the world uses Gnu Emacs, and there is pretty good software around
for using it as a Lisp top-level.  Perhaps we should give up on continuing
to improve Hemlock and instead put our effort into working as well as
possible under Gnu.

Advantages of Hemlock are the the extension language is FULL CMU Common
Lisp with all the associated debugging tools, versus the little sub-lisp
that Gnu makes available.  Also, Hemlock is well-integrated with our Lisp,
though I'm not sure we couldn't integrate Gnu just as well if we tried.
And finally, a lot of us in the group have some investment in personalizing
Hemlock in various ways, but that doesn't count for much in the real world.

Advantages of Gnu are that most of our natural "customer" base for CMU CL
is using Gnu already, and they're not going to switch to Hemlock without a
very good reason.  Hundreds of clever undergrads are out there improving
Gnu in ways that we cannot hope to keep up with, even if we wanted to
duplicate the effort.  And Gnu can be broken off and run by itself on a
small machine, without a big common lisp around.  (Yeah, someday with our
delivery mode we can hope to match that, but not right now.)  Plus, without
Hemlock, our core images might look a great deal less bloated.

Anyway, we should discuss this, poll our users, and reach some sort of
consensus on what we want to do long-term.  How we handle a switch-over, if
that is the decision, is a separate issue.

-- Scott