COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	\\M1BASL30\M2BASB30\M3BASI30
C00014 ENDMK
C⊗;
\\M1BASL30;\M2BASB30;\M3BASI30;
\F1\CJanuary 3,1979







Mr. Roger E. Vaughn
883 Lansdown Court
Sunnyvale, CA 94087





Dear Mr. Vaughn:


\JThough it has been about six weeks, I hope  you  recall
a conversation we had in Dr. Ruth E. Davis's office
on November 27, 1979. This letter is a tardy follow-up to that meeting;
the tardiness is due to the crush of holiday travel and activities, not
to a lack of interest. I painfully typed the first two-thirds of this 
letter only to leave that dreadful machine before I was finished.
At any rate, here is the letter, now produced on a semi-civilized system.

In brief, I wish to produce a comprehensive book/computing system for LISP,
much in the spirit that UCSD treated Pascal.

There are several interrelated reasons for considering this to be an
intellectually (and financially) viable enterprise.  Indications
are that LISP will be one of the major forces in the 1980's.  Though LISP
was invented over twenty years ago, it is only recently that LISP is becoming 
popular. Several forces are at work here: in the past, LISP required
very large, and therefore very expensive, computing facilities. Recent
technological developments has significantly reduced the cost of LISP computing.

Second, the computing community is becoming more sophisticated; the problems
that are being posed for solution are becoming more complex; traditional
languages supply neither the language facilities nor the programming
environments that are necessary to translate these problems into
computing terms. LISP was designed for Artificial Intelligence research --a
problem area that thrives on attacking complex, ill-specified problems.

Third, we have seen the emergence of a new force in society and computing:
the personal computer. Last year's BYTE magazine readership poll showed a very
high interest in Artificial Intelligence related topics. Since the
majority of extant AI work is written in LISP, there is a natural market
for LISP texts and related material. Unfortunately (or perhaps fortunately)
the current collection of LISP programming texts is \F3very\F1 poor. The major
offerings are \F3The Little LISPer\F1 and \F3Let's Talk LISP\F1; neither are
acceptable. They both suffer from a "cute" informal style, 
and address only  the superficial
mechanics of the language. They do not deal with the philosophy and methodology 
of programming; and they ignore the concepts of abstraction and problem solving.
These books are not appropriate vehicles for introducing LISP to the personal
computer owner. My \F3Anatomy of LISP\F1 is also not an appropriate primer;
that book presents fundamental concepts of computer science from a "purified"
LISP view-point. Little is said about the details of LISP programming; that
was intentional. The art of programming is best "done", rather than talked
about; what one needs is a primer that reinforces the programming activity.
Of course, this approach implies the availability of an appropriate LISP
implementation, and when \F3Anatomy\F1 was being written such implementations
were not available; they are still not available. The goals of this project
are to supply the primer and the implementation.

The fourth factor to be considered is related to the language named ADA.
This language is being designed for the Department of Defense as the major tool
for government software. The last government-sponsored language effort 
was COBOL. We can expect that ADA will have significant impact on
the computing community, probably
even greater than COBOL. Bill Carlson, the ARPA spokesman for the ADA effort,
was quoted as  saying the ADA is "something like LISP ... something like
Pascal". Therefore an understanding of LISP could improve one's understanding 
of ADA.
Furthermore, the  ADA programming environment team is using the LISP
programming environment  as its major model for ADA. By a "programming environment"
we mean that the programmer develops the program in a "culture" that includes
a collection of editors, interpreters, compilers, and deguggers,
all integrated in such a way that the  program --the object of study--
can be manipulated and modified even while it is being executed. Contrast this
with the traditional view in which a program is viewed as a text file 
or card deck that must be compiled, then linked, then run; if an error occurs
during the run, one must abort the computation, return to the editor, and modify the
text; then iterate the whole process. This second scheme is much like
planning a trip, beginning the journey only to find a detour, then returning
home to replan the trip, rather than modifying the original plan at the detour
and continuing.

The second style of programming is associated with "batch processing"; the
first style is "interactive". It is this interactive flavor that the ADA team
wishes to encourage in its development system. It is the interactive style
of programming
that has made personal computing popular.  The most sophisticated interactive 
programming systems built to date have been done in LISP. An interactive
style of programming is an acquired skill, best learned by experience; it
therefore requires access to an appropriate system, coupled with exposure to
materials that reinforce both interactive programming and careful program
preparation. Again, a combination of primer and system is needed here.

To sumarize the fourth point: the LISP programming style represents a view of
computing, popular to  the personal computing community and to the growing
ADA effort. This programming style encourages a non-traditional view of
computing --differing from that espoused by FORTRAN or even Pascal. Systems must
be available to allow people to experience this different approach; texts
must be available describe the programming methodology and technology
that supports the systems.

Currently, I am attacking both of these problems. My company, \F2The LISP 
Company\F1, has completed the first version of an inexpensive LISP system for the
Z-80 microprocessor. That system includes a primer written by me
as a first step towards a comprehensive text. Much has to be done, both with
the implementation and the text, before  the project is completed.

As I mentioned in our conversation, I was not pleased with the  way McGraw-Hill
handled \F3Anatony\F1. I expect to  produce my next book at least to the
point of compatibility with a commercial typesetter. Whether the matter passes to
other hands at that point is open to discussion. Several of my colleagues
have formed their own firm for publishing 
technical books (Tioga Inc. of Menlo Park);
I realize that an established firm should be able to offer more: better editing,
better distribution. Of course, that has to be balanced against the possible 
frustrations of passing responsibilities to others.

I would consider discussing a couple of possible arrangements: 1] a traditional
publisher/author agreement for my LISP book, and 2] perhaps a joint venture
for the development of both the book and the inexpensive LISP system.

I look forward to hearing from you or your colleagues. Reast assured that the
book and the system will be completed; if these enterprises can be accomplished 
in a mutually satisfactory way, I would be quite pleased.
\.
\←L\→S\←R\-L\/'2;\+L\→L

Yours sincerely,


John R. Allen
Founder, The LISP Company
18215 Bayview Dr.
Los Gatos Ca, 95030
(408) 353-3857

\←S\→L