Details

["Scott E. Fahlman" <Fahlman@C.CS.CMU.EDU>]

Martin,

Our tentative list of committee members is as follows:

Steering committee (concerned with ISO politics, etc.):

Bob Mathis, Steve Squires, Ron Ohlander, John McCarthy, Guy Steele, Dick
Gabriel.  (Basically, this is three political types, Steele and Gabriel
as liaison with the technical group, and McCarthy for added clout.)

Technical committee: Steele, Gabriel, Moon, Fahlman, Danny Boborow,
Jonathan Rees, and you.  Weinreb withdrew, since we all thought it
looked bad for Symbolics to have two members when we had to turn other
companies down.

Alan Bawden is probably first alternate; we all think that he would fit
in well, but didn't want to add both him and Rees from MIT.  Rees gives
us a formalist and a tie to the Scheme world.  We are under considerable
pressure from the Franz people to add John Foderaro to the list, but
we're trying to resist that without starting a fight.

We hope to add a couple of foreign members later.  One prime candidate
is Jerome Chailloux, if he's interested in working on this with the
understanding that things are going to be rather conservative from here
out, at least for the full Common Lisp language.  If he wants to combine
this with work on a subset or on a totally different EuLisp, fine.
Another candidate, as soon as we can establish reliable network
communication with him, is Masayuki Ida, who translated the Common Lisp
manual into Japanese and seems to be coordinating Common Lisp
standardization efforts over there.

I should mention that this is formally just an ad hoc technical
committee until such time as the X3J13 subcommittee of ANSI is formally
established, sometime this fall.  People get on that committee by
paying dues ($150/year, I think) and participating in the doings of that
committee.  This is the formal body that will establish the formal
technical committee under ANSI rules.  What we're trying to do is get a
group together to make a bunch of decisions before next fall, in the
hope and expectation that X3J13 will eventually rubber-stamp both the
decisions and the committees that were established.

I'm not sure how often elections are held for these committees.
Certainly one could resign and be replaced whenever that seemed
necessary, but I would hope that none of us would do this in the first
six months of the committee's formal existence (i.e. not before a year
from now).

The committee would have to prepare a recommendation for a language
standard to be adopted by ANSI and ISO.  We are exploring now whether
this can take the form of the Steele book, plus a set of clarifications
and changes, or whether we can use an updated edition of the Steele book
without getting into copyright problems, or whether we need some new
document written in "standard-ese".  Whatever the case, you would have
to read this thing but not write it (unless you want to).  Steele and I
will probably end up doing that, maybe with some outside help.

I'm not sure what to expect in the way of meetings.  I'm hoping that
most of the Washington/Geneva stuff can be handled by Mathis and maybe a
couple of other members of the steering committee.  I can't imagine that
there will be more than two meetings a year at which serious technical
stuff would be done.  In fact, I'm not at all sure that we'll need any,
since we've all become very adept at working things out via network.
There may be additional meetings to satisfy various bureaucratic needs,
but I doubt that you'd have to come to any of these, especially if they
have to be on particular dates in Washington.  I would hope that the
substantive meetings would take place in conjunction with other events
we're all going to or at elast would alternate coasts.  All I can say
for sure is that I'm willing to put time into this thing, but not to
travel a lot, so I'll be fighting like hell to keep that to a minimum.

The realtion to the other CL-xxx committees is easy.  They work out
proposals on various issues, but it's up to use to decide what goes
into the proposed langauge standard.  I'm monitoring all the subgroup
mailing lists, as is RPG, so nothing will take us by surprise.

The relation with the Europeans is a bit muddled right now.  Chailloux
is pretty reasonable, but if you saw the recent note from John Fitch,
you know that it is his view that the EuLisp group will develop a lisp
of their own design, drawing from COmmon Lisp but not constrained by it,
and that this will then become ISO Lisp.  The Eulisp group is probably
less than 5% of the size of the COmmon Lisp community in the U.S. and
Japan by any measure you care to choose: number of users, number of
machines, number and size of manufacturers, whatever.  But in ISO it is
one vote per country and the European standards committees tend to be
dominated by academics rather than manufacturers, so they might be able
to pull this off.

I see several possible scenarios:

1. The Eulisp effort falls apart or stalls due to internally
incompatible goals.  Since I've been on their mailing list, this looks
like a real possibility to me.  They haven't reached the hard part yet,
and have much less incentive to compromise than we had.

2. We cooperate with the EuLisp people, or maybe just those at INRIA.
They participate in the cleanup of big Common Lisp, and lead the effort
to work out subsets or multi-level specifications.  This is the solution
we all prefer.  Scheme is doing quite a good job as the clean Lisp for
academic purposes.  I'm not sure where EuLisp would fit.

3. They decide to go it alone, unconstrained by the Common Lisp design.
If they want to become an ISO standard, we agree to coexist: ISO Common
Lisp and ISO EuLisp.

4. If they insist on being the only ISO Lisp, we see who can gather the
most votes.  Maybe some of the European manufacturers will assert
themselves and force their countries to vote for Common Lisp over
EuLisp, rather than being cut off from the U.S. AI community.

5. If the Europeans win the vote at ISO, or if we just don't want to get
into a big fight over it, we simply standardize Common Lisp under ANSI
and let ISO do whatever silly thing they want it to.

I think that options 4 and 5 are pretty low probability, John Fitch
notwithstanding.

-- Scott