COMMENT ⊗   VALID 00004 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	Xerox and Common Lisp
C00006 00003	∂10-Sep-84  1434	SQUIRES@USC-ISI.ARPA 	Re: Xerox and Common Lisp   
C00008 00004	∂11-Sep-84  0625	OHLANDER@USC-ISI.ARPA 	Re: Xerox and Common Lisp       
C00013 ENDMK
C⊗;
Xerox and Common Lisp

Ron & Steve:

I talked to Beau Sheil at Xerox PARC last wednesday (Sept 5) at his
request.  The topic of discussion was how to maintain the existing Xerox
customer base while shifting towards a more standard Lisp. Sheil expressed
several concerns.

First was that Xerox did not have the resources to join in the Common Lisp
effort at the start, nor did it have the foresight to do so.

Second, the existing InterLisp user base ought not be left behind, as
would be the case if Xerox were to abandon InterLisp in favor of Common
Lisp.

Third, Sheil did not think that both InterLisp and Common Lisp could be
supported by Xerox at the same time.

Fourth, Sheil thought that the InterLisp file system style (that is, files
are ignorable side effects of user activity) and the other InterLisp
user-interface tools could not be supported easily within Common Lisp.

Fifth, Common Lisp looks too much like the Lisp sold by Xerox's
competitor.

And sixth, there is too much existing InterLisp code to easily transport
to Common Lisp.

Sheil proposed the following solution: A standards committee should be set
up to explore a subset of Common Lisp which would be easily supportable by
Xerox and which would not burden their current user base. I believe his
intention was that this subset also would be moved towards InterLisp.
Therefore, I believe his proposal is to compromise between InterLisp and a
subset of Common Lisp.

My views on this proposal are mixed. First, I think that, even ignoring
DARPA's position, the commercial world is moving with enough force to make
Common Lisp a de facto standard. Second, on the other hand, I think that
he and Xerox deserve a fair hearing on their views. If the compromise was
that a subset of Common Lisp were the DARPA standard, and if that subset
could easily be absorbed by Xerox, then I think that that is worthy of
consideration.

∂10-Sep-84  1434	SQUIRES@USC-ISI.ARPA 	Re: Xerox and Common Lisp   
Received: from USC-ISI.ARPA by SU-AI.ARPA with TCP; 10 Sep 84  14:34:22 PDT
Date: 10 Sep 1984 17:34-EDT
Sender: SQUIRES@USC-ISI.ARPA
Subject: Re: Xerox and Common Lisp   
From:  "Stephen L. Squires" <SQUIRES@USC-ISI.ARPA>
To: RPG@SU-AI.ARPA
Cc: Ohlander@USC-ISI.ARPA, Squires@USC-ISI.ARPA
Message-ID: <[USC-ISI.ARPA]10-Sep-84 17:34:27.SQUIRES>
In-Reply-To: The message of 10 Sep 84  1415 PDT from Dick Gabriel <RPG@SU-AI.ARPA>

I see no reason to weaken the Common LISP effort to satisfy Xerox.
While we should go out of our way to work with them in an attempt
to avoid loosing the INTERLISP community, I think that it is very
important that insist that they move towards advancing the
state-of-the-art. Although most of their arguments may sound
plausable, they are not necessarily hard facts. At this point
it is upto XEROX to join the new AI community that DARPA is
creating, and up to DARPA to help them without weakening its
commitment to make progress.

∂11-Sep-84  0625	OHLANDER@USC-ISI.ARPA 	Re: Xerox and Common Lisp       
Received: from USC-ISI.ARPA by SU-AI.ARPA with TCP; 11 Sep 84  06:25:33 PDT
Date: 11 Sep 1984 09:25-EDT
Sender: OHLANDER@USC-ISI.ARPA
Subject: Re: Xerox and Common Lisp   
From: OHLANDER@USC-ISI.ARPA
To: RPG@SU-AI.ARPA
Cc: squires@USC-ISI.ARPA
Message-ID: <[USC-ISI.ARPA]11-Sep-84 09:25:40.OHLANDER>
In-Reply-To: The message of 10 Sep 84  1415 PDT from Dick Gabriel <RPG@SU-AI.ARPA>

Dick, 
	I sympathize with Xerox's problem but, even were I so
inclined, I cannot subvert the standardization issue for one
company or class of users.  The fact is that a large number of
Interlisp users have converted their programs without significant
cost.  I also find it difficult to believe that a company with
the resources of Xerox cannot support both Interlisp and Common
Lisp.  That is absurd on the face of it and indicates a lack of
commitment to the AI community and products to support them more
than anything else.

I will give Xerox a fair hearing but the problem is that they are
one company trying to turn the tide of progress.  I very much
want Xerox to be a player in supporting Common Lisp and think
that it is to their best advantage to do so.  If they provided
both Interlisp and Common Lisp, their machine would be much more
attractive to a broader base of users.  In fact, I believe their
reluctance to do so shows poor business judgement on their part.
With companies like Symbolics and LMI supporting multiple
languages, how can Xerox ay that they cannot.  If we were to
accede to Xerox's point of view, we would be saying that we are
willing to compromise to meet everyone's particular needs and
requirements.  The result would be a slackening and diffusion of
the whole effort.  There may be, as you say, suffient momentum in
the commercial world to ensure a de facto standardization on
Common Lisp, but the standard would be weakened and most likely
highly differentiated amongst various users.  We don't want to go
back to the bad old days of a different Maclisp in every
computer.

Ron

P.S.  What is Sheil's position in the company.  Maybe we should
make some effort to talk to someone high up in the company.