COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	Charters
C00010 ENDMK
C⊗;
Charters
1. I'm not sure I'm ready to see a formal, legal organization take
over Common Lisp yet. I would be persuadable if there were adequate
checks and balances against random changes. Let me give an example. As you
all know, Rod Brooks and I put together a critique of Common Lisp. In it
we presented some possible criticisms of Common Lisp. The paper was
meant to stimulate thought about the resulting Common Lisp design and
about the design process. I believe that there are people who would take
that paper as a bible and proceed to implement an aggressive subset
strategy, perhaps as a means of making their old stand-by Lisp system
the `real' common Lisp. I would be the first one at the parapets to
fight against that, and perhaps Brooks would be the second. 

Therefore, I want to pursue a conservative strategy in which the current
Common Lisp committee and the members of this subgroup create a strong
document outlining how the language changes and under which circumstances
subsets are allowed  to be called `Common Lisp subsets.'

2. I believe Common Lisp will only succeed if there are excellent
implementations of it. If there are to be public-domain implementor's
kits, I want them to be high quality. I have not seen any Lisp
implementor's kit that I believe satisfies my standards. In order to
get such a kit, I think serious money will need to be thrown at the problem.

I think that one of the Universities - Stanford or CMU - could be funded
to maintain the validation suite.  I think that such an organization is the
perfect place for this role.

On the question of calling a spade a spade, why not get together a minimal
fund or assurances of free advertising which can be used to publish
notices from the Common Lisp watchdog committee about non-Common Lisps.
The idea would be this: The constitution of the CL watchdog committee would
specifiy procedures for confronting a suspected CL bootlegger. The bootlegger
would get an opportunity to explain his plans for getting a real CL together,
or else he would explain in what ways his Lisp is a subset. If he is moving
with due dilligence towards conforming, nothing would happen, except a pat
on the back and the offer of a shoulder to cry on. If the answers are
unsatisfactory, then the watchdog group would simply put an ad in the
AI magazine, SIGART, The Wall Street Journal, etc. stating: Warning,
Megabucks Inc's Lisp is not a common Lisp, it is just Lisp 1.5 with arrays.

3. Lisp's affiliation with AI is touching, but I'm not sure I would want
an AI organization as the watchdog over CL. Especially AAAI. Do you remember
the last Lisp conference? AAAI gave us many assurances that AAAI would not
overlap the Lisp conference. AAAI started their session on tuesday morning,
I believe. They avoided overlapping by having their opening session on
thursday morning. (As an aside, as chairman of the next Lisp conference,
I considered two alternatives: 1) the Lisp conference could hold its
closing ceremonies immediately after the opening ceremonies or 2)
the Lisp conference could be separate. I chose the latter).

Perhaps a non-profit corporation could be set up to do it to help pay for
administrative matters, but, again, I would not support it until there
was a strong CL constitution.

4. I think that the government and the companies should pay for the
adminstration of the CL watchdog. 

5. I think that the CL constitution ought to state the CL design
principles.  For example, the constitution ought to state that CL is a
lexical, multiple-value, etc. Lisp. Then, there ought to be procedures for
undoing those decisions, but doing that ought to be very hard. The minor
points - like EQL instead of EQ for CATCH tags - can be left up to the normal
functions of the watchdog group.

I am, frankly, worried about the power politics involved in all of this.
I know it might sound silly, but I tend to want to model the organization
of the CL world after organization of the US government, *but* only to
the extent of the idealization of a constitutional republic. 

Let me illustrate my concerns.  We could easily posit that companies would
get representation proportional to their size, the size of their
investments, or the percentage of their capital devoted to Common Lisp
activities.

Under any of these schemes, some group would get an unfair amount of
power. Under the first scheme, IBM probably (or Futjitsu) gets the most,
under the second, probably TI gets it, and under the third Lucid gets it.
Now I don't mind other people calling shots, but I do mind if one those
groups, say Lucid, called the shots and it was fundamentally wedged in its
judgement on some point.  How could that group be prevented from forcing
its will on the rest of us? A strong statement of what must remain in CL,
at least in principle, would go a long ways towards that. A second way to
go a long way is to have a distinguished panel of Lisp wizards as a court
of appeal or advice.

6. I don't advocate democracy, and my model reflects that.

			-rpg-