[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
iso work on Lisp
- To: MATHIS@USC-ISIF.ARPA
- Subject: iso work on Lisp
- From: "Scott E. Fahlman" <Fahlman@C.CS.CMU.EDU>
- Date: Thu, 15 May 1986 23:58 EDT
- Cc: cl-steering@SU-AI.ARPA
- In-reply-to: Msg of 15 May 1986 22:32-EDT from MATHIS at USC-ISIF.ARPA
- Sender: FAHLMAN@C.CS.CMU.EDU
I'm not sure how it is supposed to work in ISO when two groups want to
standardize the same thing in two incompatible ways. Do we both develop
proposals and then fight it out, or are we supposed to fight it out now?
Or is the convenor supposed to try to reconcile the two groups? What
exactly does a New Work Item mean?
Our position, I think, is that Common Lisp as described in ClTL is
already a de facto standard, and we intend to propose an official
standard that reflects this. Any changes will be considered in the
light of their effect on existing user code and implemenations;
incompatible changes will only be adopted if the benefits clearly
outweigh the costs. From what they have said in the past, the Europeans
intend to propose lots of gratuitous incompatibilities -- they didn't
get to play as Common Lisp was being designed, so now they feel free to
change everything they don't like, such as the type of NIL.
I think that these are fundamentally incompatible goals. A compromise
that would let them make just a few incompatible changes is as bad as
letting them redesign the whole language; the existing Common Lisp
community wouldn't buy gratuitous changes at this point just to make the
language more elegant in some eyes. The only other kind of compromise
that I can see is to let us do "ISO Common Lisp" while they try to
develop an "ISO EuLisp", but they seem to reject this since they want to
do "ISO LISP" and they have stated that the current Common Lisp should
not be standardized since it offends their sense of elegance. So unless
the Europeans relent, I think we have to go through whatever process is
invoked when irreconcilable differences arise in the standards process.
If we lose, we've still got ANSI, I guess.
If it comes to a shootout, they've got a few experts, but we've got
heavier ones. I'm not sure how we'd stack up in terms of publication
count, but we certainly have more practical experience by any
conceivable measure. We've also got a large and growing user community
and all of the major U.S. companies, which ought to count for
something. The Europeans are still trying to decide how to define what
a "Lisp" is, and they have nothing remotely resembling a complete
design, let alone implementations. So far, they've postponed all of the
kinds of decisions that lead to real disagreements. On the other hand,
the Europeans probably have the votes of more countries than do right
now, though I suppose we could start recruiting in the third world.