[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Standards
- To: Mathis@USC-ISIF.ARPA
- Subject: Standards
- From: "Scott E. Fahlman" <Fahlman@C.CS.CMU.EDU>
- Date: Sat, 14 Jun 1986 15:49 EDT
- Cc: cl-steering@SU-AI.ARPA
- Sender: FAHLMAN@C.CS.CMU.EDU
Bob,
I recieved your packet of material obtained from the Eulisp group.
Thanks.
One point they make is that the creation of a separate Common Lisp
committee under ANSI is in violation of ISO rules and that the British
stadards organization is upset about this. Is this true? Is it
serious? I have found it reassuring to believe that, whatever stupid
thing ISO does, we would at least have an ANSI standard for Common Lisp.
If they can prevent that, while seizing control of the ISO machinery,
then we're back to having a completely informal standard for Common Lisp
that everyone follows, while there is an official ISO standard that is
supported by six people (and 20 countries) worldwide.
Obviously there are a few people in that group have are personally quite
bitter about Common Lisp passing them by. Stoyan, Padgett, and Fitch, I
believe. There were some beautiful examples of venomous writing in
their internal document, all aimed at Common Lisp. They make excellent
use of the Brooks and Gabriel critique as a weapon against us. I don't
know what one does to assuage such intense feelings, except to
reciprocate with as much reasonableness as we can muster and hope that
they dissipate over a period of a year or two.
They repeatedly level three criticisms at the existing Common Lisp:
1. It is big.
2. The current spec is imprecise.
3. It ignores all Lisp developments outside the Maclisp world.
Their solutions:
1. Multi-level definition.
2. Level-0 is to be specified by formal means, probably denotational
semantics, and the higher levels are somehow supposed to inherit
precision from that.
3. Encouraging a more widespread discussion of of Lisp standardization.
My reactions:
1. Yes, it is big. This isn't a problem, in my view. If it were
a problem, it should be solved by one or more true subsets of Common
Lisp, rather than by starting over.
2. Their solution won't work, or at least it is something that has never
worked in the past. The best we can do is to take a
well-tested, widely used Lisp and clean up the problems that
people have discovered.
3. It is demonstrably false that Common Lisp ignored everything outside
the Maclisp tradition. Steele was the co-inventor of Scheme, and we
got input from many other groups. What they are really saying is
"WE weren't consulted about this." That is both true and
regrettable, but throwing away Common Lisp and starting over is
hardly the appropriate remedy, now that Common Lisp has become a de
facto standard supported by every company that matters.
We should have a joint meeting with the Eulisp people in Boston to see
if there is any acceptable way to settle this.
-- Scott