COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	\\M1BASL30\M2BASB30\M3NGR25\M4NGR20
C00010 ENDMK
C⊗;
\\M1BASL30;\M2BASB30;\M3NGR25;\M4NGR20;
\F1\CJan 23,1980



Mr. Phil Villers
20 Whits End Road
Concord MA 01742

Dear Mr: Villers:

\JVic Scheinman suggetsted that I contact you about my company's
plans and expectations.  The next  paragraphs outline the current
state and suggest how
ComputerVision or you might aid and profit from the situation.

First, there is an existing market for LISP machines. Besides the
traditional applications in Artificial Intelligence (including
natural language understanding,
medical diagnosis, chemical and geological analysis, and of course, robotics), 
LISP has become quite successful
as a general implementation language for complex systems (including
algebraic manipulation
systems like REDUCE and MACSYMA, computer-aided design, data-base
systems, and graphics languages; even the MIT LISP machine was implemented
in LISP).

At the high-end
of the LISP machine market one finds either special purpose hardware or 
re-microcoded
 commercial units. In the lower end of the spectrum, microprocessors
are being utilized (even eight-bit processors do suprisingly well).
One difficulty is that there is an insufficient supply of qualified LISP
implementors; I know of at least three projects currently stalled because they
cannot locate qualified staff.

Second, the demand for this kind of a machine is growing.
ADA, the new DOD language, has been described  by its sponsor as "something
like Pascal,... something like LISP".
Further, the ADA team that is developing
the "ADA programming environment" will be using the LISP experience as their model.
LISP has a rather distinctive view of program development --to a large extent,
due to its AI ancestry. This "view" mainfests itself both as an "interactive
style" of programming and as a collection of hardware and software "tools"
--the environment-- 
that support  and surround the actual LISP language. 
Since ADA is expected to deal with problems as complex as those
investigated by AI, these aspects of LISP  are very important to the ADA team.

Also,
the LISP applications market is just beginning to flower.
The performance of AI programs
has improved immensely in the past few years. AI programs
called "expert systems"  can now perform comparably to 
specialists in areas as diverse as organic chemistry, 
medical diagnosis, and geology. Unitil recently, AI programs
required large expensive conputing systems; technology has
changed that. The computing power of the new microprocessors
is sufficient to support an elegant LISP implementation;
the cost of these components is sufficiently low that this  
implementations can be "mass produced". With these inexpensive systems one
can then package many of the AI programs and
apply
these programs to society's needs. 
One of society's "needs" is the personal computer; one can imagine
many applications of AI ideas to this domain to make computing devices
more responsive to naive users. There is already a strong LISP interest in the
small computer area as the last BYTE readership poll shows; 
that interest was one of the  reasons that BYTE asked me to 
guest edit the August 1979
special issue on LISP. Further, I have established a
"LISP users/interest" group whose membership list keeps growing, even though
LISP is only now becoming available on personal computers.

Finally, I have been approached by several possible clients to develop LISP
on their machines. I have implemented a rich subset of  the LISP found on 
MIT's LISP machine. My project was done for Cromemco on their Z-80 --a
project somewhat akin to building a scale model of the Eiffel tower
in a perfume bottle.  Rather than continue my comapny as a supplier of contract
software, I wish to develop LISP machines as completed hardware/software
systems.  Given the LISP tools, I can then begin to exploit the
AI applications. 

I would like to establish The LISP Company as an independent organization,
to supply LISP and LISP-related products. Such a venture requires capital.
My initial goal was to develop the necessary funds slowly as TLC grew.
However, it appears that such growth will not be rapid enough to capture
the developing market. For many years, Japan has been active in AI research;
there are at least six current LISP machine projects underway in Japan
now. In this  country, LISP has been more difficult to sell to industry;
indication are that this attitude is beginning to change.
My LISP book (now being translated into Japanese), and the special issue
of BYTE, as well as TLC's LISP have had very positive response.
Rather than lose this momentum to others, I am now
pursuing external funding. 
If the idea of investing in TLC appears attractive,
please give me a call and  I can supply more details.
\.
\←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