EuLisp

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

I note that some cracks seem to be developing within the Eulisp group,
and that they are nearing the point where the differences in goals
between the formalists and the hackers will begin to manifest themselves
in major technical disagreements.  It looks like the move of the May
meeting to Erlangen was the only way to keep Stoyan on board.  It will
be interesting to see if the June meeting comes off, and whether they
are still working smoothly together by then.

With respect to Eulisp and the Eulisp people, I think that the following
points should be raised with them fairly soon.  It is best to approach
this though Chailloux, as he seems to be more interested in Common Lisp
than the Padgett and Fitch.

1. We have told them this before, but we should reiterate our view that
it is necessary to standardize something reasonably close to the current
Common Lisp under ANSI and, if possible, under ISO.  By "reasonably
close", I mean that we must recognize that many Common Lisp
implementations and a growing body of user code and training materials
exist already or are in preparation, so in the definition of the
standard we must not make any incompatible changes unless the benefits
very clearly outweigh the costs.  In this process, we do not feel that
we are free to start from scratch and reconsider all of the old
decisions.  We certainly do not view the current Common Lisp as perfect,
but it has attained the status of a de facto standard in the U.S. and
many other parts of the world, and orderly progress demands that we make
this standard explicit and official and that we clean up the current
ambiguities as best we can.

2.  It is our intention to develop a cleaned-up language specification
for the full language, and submit this to ANSI and then to ISO as a
proposed standard for ANSI/ISO Common Lisp.  We do not view this as
casting the Common Lisp spec in stone forever, nor do we wish to
preclude the emergence of an ISO standrd for some different Lisp or for
an official Common Lisp subset or set of "layers".  But for all of us,
the first priority is developing a usable standard for the full
language, and we do not currently view a layered approach or a
definition using formal semantics as the quickest or best way to attain
that goal.

3. We recognize that some members of the Eulisp group want to develop a
"post-Common" Lisp that would be different in many ways from the
currently defined language.  Others appear to want a mulitple-level
specification, with the most complex level corresponding closely to to
the current Common Lisp and with the lower levels being subsets.  The
simplest levels may be definable by formal mathematical methods.  We
respect these goals and wish you well, but we hope that such activities
will not preclude an ISO standard for something close to the current
Common Lisp with a specification that discusses only the full language.

4. If the Eulisp group or any individual members wish to contribute to
the standardization effort outlined in points 1 and 2 (even while
pursuing your own activities as discussed in point 3) we would welcome
such participation, either through the open discussions on the Common
Lisp mailing list, or perhaps by adding a Eulisp person to our technical
and/or steering committees.  We would like to discuss this committee
membership if there is serious interest.

5. If you feel that you cannot subscribe to our goals and plans, we
understand, and wish you well.  Even if we disagree on the best form for
a standard, we hope that the gorups can remain in close contact and can
learn from each other's efforts.

-- Scott