COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002				  May 9, 1981
C00031 ENDMK
C⊗;
			  May 9, 1981













Mr. Tom Dornback
1900 N. Austin
Chicago Ill, 60639

Dear Mr. Dornback:

I enjoyed talking with you about TLC-LISP and my plans for LISP in general. This
is the first  opportunity I  have had  to follow  up on  our conversation.   I'm
enclosing some  information about  TLC-LISP; furthermore,  this letter  suggests
some opportunities to exploit LISP and its applications.

Those things involve:

⊗ Competition in the LISP market. 

⊗ TLC-LISP in a learning environment: the spring Santa Clara  University course, 
  the summer Western Institute in Computer Science course, and the SCU Humanities 
  and Sciences program.

⊗ The Commercialization of AI and TLC-LISP: Applications of the existing Z-80 LISP 
  in single and multi-bank configurations.

⊗ Possible strategies for marketing LISP.

⊗ The Future: An extended TLC-LISP for the 16-32bit machines.




			  Smalltalk, LOGO, and LISP   

First, direct competition  with TLC-LISP  in the micro  LISP market  is easy  to
discuss: it's poor. The only micro LISP used by the professional LISP  community
is TLC-LISP: MIT, BBN, and the Air  Force and Navy, all have projects  utilizing
TLC-LISP. My difficulty is  a non-existent marketing  effort; current sales  are
mainly the result of word-of-mouth contacts. With these constraints, I have sold
approximately 100 systems, and an equal number of the  TLC-LISP manual.

TLC-LISP is a sub-set of the latest LISP developments at MIT; other micro  LISPs
are either variants of the original LISP1.5 (1959) or LISP1.6 (written by me  in
1965). Both of these efforts are substantially out-of-date.  So, given marketing
support and  the  testimonials  from  the  current  TLC-users,  TLC-LISP  should
demolish the competition.  What other languages are contenders for the LISP view
of computing? Smalltalk and LOGO.

Last  September,  Xerox   announced  the   release  of   Smalltalk-80  to   five
manufacturers, beginning efforts to bring  the Smalltalk family of languages  to
the mass market.  Smalltalk's major attractions  to the user  are: (1) a  highly
effective graphical user  interface, and (2)  a user-oriented programming  style
called "object-oriented programming". These two characteristics combine to  give
Smalltalk users a  flexible medium  (almost "non-programming"  medium) in  which
they can pose their problems.   Problem  domains range  from animation to office
automation.  The effect of Smalltalk is expected to be massive.

LOGO, as  described  in  Seymour  Papert's  new  book  "Mindstorms:   Computers,
Children, and  Powerful Ideas",  has been  an exceptional  vehicle for  teaching
primary-grade students "how to think".  LOGO is Smalltalk's immediate parent. It
was developed  at  MIT in  the  late 1960's,  introducing  the idea  of  "turtle
graphics", and implementing the  rudiments of object-oriented programming;  both
hallmarks of  Smalltalk.  So  LOGO  offers a  direct  counter to  the  Smalltalk
challenge.   However,  it  is  limited  in  scope,  addressing  itself  to   the
educational domain rather than commercial quality applications. A good LISP  can
do better than either LOGO or Smalltalk.

The appropriate view is to consider LISP as a "high-level" assembly language for
describing complex  problems.  In  this  view,  then,  Smalltalk  and  LOGO  can
accurately be described as higher level languages which could be  implemented in
LISP. Indeed, many versions of LOGO (including the original one) are implemented
in LISP.   The object-oriented  style and  the graphical  interfaces are  simply
special-purpose  applications   of  LISP   code.  This   has  been   effectively
demonstrated on the MIT LISP machine.

Besides being the  parent of  many of the  LOGO/Smalltalk ideas,  LISP offers  a
degree of maturity and practicality not found in either Smalltalk or LOGO.  This
breadth includes  Artificial Intelligence  applications as  well as  educational
benefits.

In the  educational area,  LISP  shines.  For  the underlying  mechanisms  --the
reasons-- behind the "flash" of  the Smalltalk/LOGO programs are easily  grasped
in terms  of the  semantics of  LISP.  This  explanatory aspect  is critical  in
education, and has not been handled adequately in either Smalltalk or LOGO.



			  The Commercialization of AI

The AI  market is  broad, both  in the  sense of  professional applications  and
educational implications.  Major corporations --Schlumberger, Standard Oil,  TI,
CDC, IBM, and  HP for  example-- are  investing heavily  in AI.   These are  not
simply research efforts;  they are  directed to the  substantial development  of
commercial products.  These  applications include intelligent  CAD systems  that
understand  design  rules;   CAI  systems  that   understand  the  topic   under
investigation; "expert  systems"  that  can  operate  at  the  expert  level  in
specialized  domains  like   chemistry,  genetics,   medicine,  and   structural
engineering; intelligent  control  systems  for  robots;  large-scale  algebraic
manipulation systems;  and  natural language  interfaces  to complex  data  base
applications for business data processing.

The immediate impact  of this  is a  short supply  of trained  AI personnel  and
therefore increased demand at  the college level  for AI-related courses.   From
beneath this surface will spring a  general cultural awareness of, and  interest
in, AI-related topics. For  example, four publishers have  asked me to write  an
introductory-level  LISP   text;  and   a  group   of  us   organized  a   major
session/demo/tutorial on  LISP  and  related  areas at  the  recent  West  Coast
Computer Faire. This was a great success; we had 500-600 people who attended the
all-day LISP  seminars. So  an audience  is  there; one  only need  exploit  the
curiosity.

The point is that AI is a  growing market --two billion dollars by 1990,  claims
Business Week-- and  a good portion  of that  work will be  done on  traditional
micro processor based machines rather than special LISP machines.   Furthermore,
this is a market that no language other than LISP will impact for many years.

In a combination of  AI and education, the  US Navy is staffing  a large R&D  AI
effort. One aspect of  their program is an  application of Intelligent CAI  that
will utilize several  thousand micro-based machines  running LISP.  One  segment
--called the STEAMER  Project-- of  this effort has  recently purchased  several
Z-80 systems solely for  the purpose of running  LISP.  These systems are  being
utilized to "off-load" parts of a LISP-based simulation model of the dynamics of
a ship's  steam  boiler  plant. The  combined  system  is used  to  train  naval
personnel in the care and maintenance of the steam plant.  Though parts of  this
work can be done effectively on Z-80 class machines, the ultimate target of  the
Navy work is, of course, a more powerful machine and a more powerful LISP.

In particular, the group that funded the UCSD Pascal effort wants me to  develop
an extended version  of TLC LISP  for 16-bit  machines.  That is,  they will  be
funding me to "do for LISP, what  UCSD did for Pascal".  Considering the  Navy's
commitment to AI,  and the effect  their Pascal effort  had on micro  computing,
this TLC project will have a substantial impact on the AI field.





			      Educational Movement

AI applications are  only one thread  of the LISP  fabric. Another major  impact
will be "nose-to-nose" with Smalltalk in the educational market, for LISP offers
several advantages here.   At the level  of applications, one  can easily  argue
that a syntactically "sugar-coated" LISP  with a graphics interface will  supply
the younger learner  with much, if  not all, that  Smalltalk supplies. 

Where LISP  comes  into its  own  in the  educational  market is  its  basis  in
mathematics and computing principles; one can use LISP to explain the principles
that underlie  the  phenomena  that  one  experiences  in  Smalltalk  and  LOGO.
Smalltalk and LOGO  are to engineering,  as LISP is  to mathematics. This  added
leverage gives LISP inroads into intellectual disciplines that are not open to a
more superficial language. For example, one can give a firm and intuitive  basis
for modern computing, based on simple LISP-like constructs.

At MIT Dr. Harold Abelson  and Dr. Andrea diSessa  have instituted a program  to
revise the undergraduate  mathematics and physics  curricula using a  LISP-based
approach.  Their project, co-sponsored by MIT and ARPA, illustrates two  points:
first, MIT's committment to LISP --LISP has long been the major language used in
their introductory software course, and now is to be used to revitalize the more
traditional  undergraduate  curricula.    Second,  ARPA's  sponsorship   denotes
awareness that the growth of AI will require substantial increases in the number
of AI personnel. The MIT program will do much to improve education and  increase
the general awareness of LISP ideas.

At  Santa  Clara  University  I  have   instituted  a  program  to  revamp   the
undergraduate computer science curriculum  using LISP ideas  as the base.   This
program includes  an interactive  programming lab  as an  integral part  of  the
educational experience.  This  pilot study began  this spring term  and will  be
exported  to  other   universities  as  rapidly   as  possible;  several   other
universities have expressed strong interest.   We will also expand this  program
into a continuing  Humanities and  Science faculty workshop;  this program  will
develop a campus-wide course to  introduce computing to the student-body.   This
program will begin this fall in the Humanities school at Santa Clara University.
Futhermore, I am  actively pursuing  publicity for  these efforts  in the  local
community; it is important  that people realize there  is an alternative to  the
dull, boring, and deadly traditional approaches to "computing and society".

My ultimate goal is  to move a version  of this integrated program  --LISP-based
text and  computing  laboratory-- into  the  high schools  to  revitalize  their
mathematics and science programs,  while giving a more  accurate view of  modern
computation than that presented in the traditional BASIC model.

The potential,  both  in  financial  and intellectual  terms,  for  an  improved
educational offering is immense.  There is increasing awareness that the current
uses of  computing  in  education  are  inadequate.   A  recent  joint  NSF  and
Department of Education task force reported grim predictions for the educational
output on the US technological front;  they suggested a Presidential council  to
guide the refurbishing of our mathematics and science programs. Recent issues of
Computer magazine, Communications of the ACM, BYTE, and U.S.News, have addressed
the same problems: current science education is a disaster.

In the July 1980 issue of  BYTE, Dr.  Arthur Luehrman suggests a  billion-dollar
equipment market and $50 million annually in sales of texts for computer-related
secondary education.   It was  in this  issue  that Lou  Frenzel of  Heath  also
addressed the educational applications of micro computers.  As discussed  above,
I definitely support a  self-paced computer-supported educational tool.   Rather
than pre-programmed instruction, I suggest that many of the techniques developed
by the AI community can  be applied to CAI  to result in intelligent  "tutoring"
systems --a combination of CAI and the expert systems work. Since almost all the
expert systems work is written in LISP,  a LISP-based product line is a  natural
starting point for these applications.

The applications and  educational market  will be there  in full  force for  the
16-bit machines, and Smalltalk's entry will only enlarge the wave.  LISP  offers
an  elegant way to finesse the Smalltalk  phenomenon, and TLC will be the  force
behind micro-based LISPs.




			    LISP Course for Industry

On the technical side, I am offering  a LISP class for the Western Institute  in
Computer Science (run by  the people who  used to run the  UC Santa Cruz  Summer
Institute) to be held this summer  at Santa Clara University.  This course  will
draw a substantial number of people, active in applied AI; the manufacturer  who
could demonstrate an inexpensive, powerful  LISP system at this Institute  could
reap substantial benefits.  





		       Extended Memory and New Processors

The key, for  applications and effective  education, is LISP  in an  inexpensive
setting that is compatible with the parameters of interactive programming.   The
initial offerings in this market can either  be the single bank version, or  for
more extensive applications, the  multi-bank TLC-LISP.  In either  configuration
the LISP runs at about 1/3 of a KA-10, when utilizing a 4MHZ processor.  In  the
multi-bank configuration, the system  gives an effective  memory space of  about
1/3 of a  PDP-10. Given an  appropriate display interface,  such a system  would
offer the  LISP programmer  a desk-top  LISP  machine, equivalent  to 1/3  of  a
million-dollar installation, for about $5K!



				   The Future

The long-range solution is an integrated system running on the newer processors,
offering larger address space  and more flexible  display systems.  The  initial
Navy grant  specifies  the  MC68000,  but  I  have  other  processors  in  mind.
Requirements for interactive programming  as defined by  LISP (or Smalltalk,  or
LOGO) are  more  stringent,  requiring careful  consideration  of  display,  and
virtual memory design.   I believe  that, working  together on  the Santa  Clara
project,  Heath  and  The  LISP  Company  can  cooperatively  develop  a  strong
educational  program  and  a  machine  that  will  please  the  growing  AI  and
educational community.


				    Summary

As you can see, mine is a broad plan whose success requires more support than  a
small company like mine can supply. I am not the University of California,  MIT,
or even  Santa  Clara  University  for  that matter;  neither  am  I  the  Xerox
Corporation.  I  am  convinced that  my  plan  is sound,  both  financially  and
intellectually.

The most pressing problem is the Santa Clara offering.  I need to locate support
immediately to assure that  the Interactive Programming  Laboratory is in  place
for the fall.  The most critical problem now is the acquisition of machines  for
the programming laboratory.


			    A Suggested Arrangement

My two  main concerns  are: (1)  effective marketing  of the  existing  TLC-LISP
offering, and (2) support  to develop the products  for the 16/32-bit  machines.
As I mentioned in our conversation, Cromemco  has not exploited its lead, and  I
as TLC do not have staff or finances sufficient to mount and sustain a marketing
drive. The product has significant potential in its educational applications;  a
comprehensive text  and lab  manual  is needed  to  complete that  package.  The
outgrowth  of  the  Santa  Clara  work   will  accomplish  that.  The  book   is
well-underway, and the  manual will  develop as a  much revised  version of  the
TLC-LISP manual. I would  be pleased to consider  an agreement linking  TLC-LISP
with the educational packages that I am developing.

The word-of-mouth acceptance  of TLC-LISP as  an AI product  speaks well of  its
ability to support AI applications.  I  need to develop a better user  interface
--editor,  debugger,  and  compiler,  plus  simple  language  extensions.    The
difficulty arises in  that I  cannot do all  these things  simultaneously; as  a
result, only the most pressing problems  get addressed.  I need assistance.   If
you are interested  in pursuing a  marketing strategy for  TLC-LISP, I would  be
most anxious to aid your  staff; I can suggest  ideas, rough copy, and  probably
supply quality "testimonials"  from well-known personalities  within the AI  and
personal computing sectors.

So my  first  major concern  is  the exploitation  of  the Z-80  potential:   an
extraordinarily versatile piece  of software on  inexpensive, widely  accessible
systems, major importance being the  educational applications, with AI being  an
important but secondary, market.

I would also consider linking the  Z-80 with your mentioned 16-bit  development.
That is, my  second major concern  is to enhance  the systems --educational  and
AI-- for more  comprehensive applications; both  intelligent CAI and  artificial
intelligence systems can utilize more power than  is available on a Z-80.  I  am
most anxious to  obtain quality systems  based on the  newer processors.   These
systems should integrate the processor with a high bandwidth display system, and
though initially  expensive, these  systems  represent product-lines  that  will
become "mass-market" items in the future.

My only viable option is to  obtain these systems from an existing  manufacturer
--I do  not  wish  to  attempt to  establish  yet  another  manufacturing  firm.
Furthermore, it is unacceptable  to wait until such  machines are announced  and
available; the software effort must be completed before that time.  Thus I would
suggest an arrangement whereby Heath would market the existing TLC product, help
support the  educational  development  at  Santa  Clara,  and  supply  prototype
hardware for the  next generation of  machine. In return  Heath would obtain  an
educational  package  that  would  establish  your  company's  position  in  the
computing/computer literacy field.

If you  would like  further  information about  any  of these  projects,  please
contact  me   through   one   of  the   following:    Santa   Clara   University
--(408)984-4611, TLC --(408)353-2227, or home --(408)353-3857.



						    Yours sincerely,


						    John R. Allen
						    The LISP Company.
						    18215 Bayview Dr. 
						    Los Gatos Ca, 95030
						    (408) 353-2227


Enc
A Rationale for EECS129
Story for the SCU newspaper
TLC-LISP Manual
TLC-LISP Flyer