COMMENT ⊗   VALID 00004 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	\\M1BASL30\M2BASB30\M3BASI30
C00018 00003
C00037 00004
C00039 ENDMK
C⊗;
\\M1BASL30;\M2BASB30;\M3BASI30;
\F1\CApril 15, 1981











John Bigbee
H.E.V. Co. Mfg.
2663 Saturn Street
Brea, CA 92621


Dear John:

\JI apologize for not getting in touch with you sooner, but last week was 
 a disaster. A government contract that was supposed to get money in
my hands by March 1, is still tied up in red tape.
A multi-year LISP project with TI  has gone into an indefinite holding 
pattern; and,
most discouraging, my course --an extended version of the Faire Tutorial-- 
and I are the target of a  political struggle  in the
Engineering School at Santa Clara University.
All  in all, a rather poor week.

Somewhat better news this week:  the Humanities School at Santa Clara has
requested  a  version of the course. They wish to offer it as  a
Faculty Workshop and  as
an early undergraduate computer literacy/culture course 
in the fall. Of course, when
I say "literacy" and "culture", I \F2mean\F1 "literacy" and "culture",
not the superficial slop that gets passed off now. My current (controversial)
offering in the EECS department is the model: a cross between 
Doug Hofstatder's \F3Godel, 
Escher, Bach\F1 and Seymour Papert's \F3Mindstorms\F1; it seems to 
irritate some members of the engineering
faculty  because I suggest that deep computational ideas can be 
made (a)-fun and (b)-accessible to  those who don't know Fortran/Basic/Pascal.

I'm working the material into a book to be used in conjunction with
an interactive programming laboratory. This package will be
available at San Jose State and San Francisco 
State, as well as Santa Clara. I hope for wider
exposure at the college level this fall, if I can 
 organize the funding. Though it will
take several more years, the ultimate target of this
program is the high school system. As LOGO attacks brain-death in 
primary grade mathematics, so too a LISP dialect offers hope for rejuvenating
high school mathematics programs.  This was the subject of the
\F3Bankruptcy of Basic\F1 paper. 

As with the LOGO effort, a robust interactive programming environment
is a necessary component. Furthermore, the cost of such systems must be 
within the grasp of  school systems. My proposed system at Santa Clara
involves a local network of  personal machines, sharing mass storage.
These were to be Z-80 based systems with memory-mapped displays, giving
a 40x80 character space in which Smalltalk-like window-based computation
could be performed. Unfortunately, I could not raise funds for this. So,
currently we share two Apples with San Jose State, and  soon will utilize
a Model II TRS-80 that I managed to scrounge. Though its display area
is small, the Model II has several advantages: it supports several banks
of memory,  is ubiquitous, and relatively inexpensive. TLC-LISP can utilize
multi-banks to increase the address space to about 1/3 of that available
on a DEC-20. Since the speed is also about 1/3 of a KA-10, this configuration
offers the possibility of a very substantial desk-top LISP processor
for about $5000. This multi-bank version has attracted attention in the
professional LISP community, but it is the less expensive single-bank version that
is more attractive in the general educational market.

Unfortunately, my approach to education requires a cooperation of cultures: the 
technical development of computational equipment, and the development and
publication of curricula that uses the technology in a creative fashion.
The LOGO  project is the only model that even comes close to fitting
my conceptions. Being on both sides of the fence I end up getting neither.
Texas Instruments seriously considered the plan last fall; they avow
corporate interest in LISP, AI and education, but seem completely unable
to make decisions.
Addison-Wesley, McGraw-Hill, Holt Reinhart and Winston, and Byte Books
are interested in the book part of the venture, but are neither willing to help
finance a major educational campaign nor support the necessary development of the
computer-based tools.
I strongly believe that the kind of
intelligent literacy, represented
in my approach, will have substantial intellectual, as well as financial
rewards. 

The pending government grant, mentioned in the first paragraph, potentially
addresses  both the educational and technical areas:
it comes from the group that sponsored UCSD in their Pascal development
effort --the Naval Personnel Research and Development Center. 
STEAMER 
--the topic of Bruce Roberts' presentation at the Faire-- is 
funded by the NPRDC  as part of a larger  program
to develop intelligent tutors to aid in the training of Naval personnel.
Initial NPRDC estimates projected 1000 LISP machines per year for 4-6
years; these LISP machines being 16-bit micro-based systems. The Navy
grant to me is to initiate the development of TLC LISP on the 
Motorola 68000. Indeed, the grant will come, but until it does my finances
are woeful. Furthermore the grant itself is not sufficient 
for TLC to reach  "critical mass" and develop the
\F3de facto\F1 standard LISP; the grant will cover one 68000-based
computing system  
 and one full time employee. Hardly sufficient to capitalize
on the  lead that NPRDC has given me.  

Even more anxiety  befalls me,
in that until I can get the money, I cannot order appropriate hardware.
Besides utilizing the hardware on the 68000-version of TLC-LISP,
I have also been promised
prototype boards from
National Semiconductor for  their 16032 processor. That processor has
has several advantages over the 68000 for LISP-like applications.
I have stirred up interest in LISP at National, but have not
 stirred up any  funding --yet. I find it intriguing that National's
chip is second-sourced by Fairchild --a Schlumberger affiliate, and
Schlumberger has a substantial  investment in AI applied to
oil exploration (the Prospector program, in particular).
Futhermore, there is the possibility of putting Smalltalk-80 on the
National machine; Xerox is interested in expanding the Smalltalk arena
and a machine that's good for LISP will be good for Smalltalk.

Though the  68000 or 16032 systems will be outside of the
inexpensive educational market for some time, they \F3will\F1 be attractive
to those people actively pursuing the commerical applications of Artificial
Intelligence. These people currently have to rely on machines in the $100K
class. A professional LISP system on one of the large address-space micro-processors
will do very well.

Related to the "high-end" technical education,
I am teaching an intensive week-long course in LISP this summer, 
 emphasizing more of the interactive computing aspects in the area of large 
Artificial  Intelligence projects. This LISP course
will attract people in industry who are actively applying Artificial
Intelligence techniques in commercial applications. We could expect significant
leverage in future sales if we could announce the projected availability of a
modest cost ($30-40K) professional LISP system.

In this letter I have touched on several areas: the broad problem of 
general quality education about computational ideas
(computer literacy, social issues, and dispelling of phobias), the more focused
issue of quality technical education (computer science  and mathematics),
and the intensely technical area of the education of professional craftsmen
(large-scale AI applications). Technically, TLC is able to address all of these
issues; realistically, I need a substantial base in hardware, software and
financing, in order to make the necessary impact.

Resource requirements are substantial for projects like this.
The newly formed company to market and develop LOGO, has opted for an
LMI LISP machine; TLC should have no less. Further equipment  representing
target machines for the development effort will also be required. These
should be inexpensive commercially available systems for the high-school
market, medium-cost, higher functionality systems for universities, and
high-end  systems for the commerical market. Target hardware therefore
runs the gamet from Apple to Apollo. 

Staff is also a critical area; there
is a general scarcity of quality LISP talent. Several LISP implementation
projects are languishing because of this shortage. Fortunately, given funds,
I believe I can attract several bay-area LISP experts. Furthermore, on the
basis of my experience with TLC, I believe that it is important  to include
a person with strong business/financial backgound. All members of such a 
team should support the global goal --quality educational and technical products.
Given this basis  of education and tools, it then makes sense
to discuss the application of these tools to 
the design and marketing of intelligent systems. But that's another letter.

The intent of this letter is to give you a bit of perspective on what I am
trying to accomplish with The LISP Company. 
I am doing this because I am  interested in investors who 
believe in what I'm doing.   I believe the potential is enormous.

In the accompanying documents, I supply a bit
more detail on the topics addressed above. I hope that this --sometimes
redundant-- material
supplies sufficient background on my goals and activities, such that we can discuss
possibilities for investment. 
\.

\←L\→S\←R\-L\/'2;\+L\→L

Yours sincerely,


John R. Allen
18215 Bayview Dr.
Los Gatos Ca, 95030
TLC         (408) 353-2227
Santa Clara (408) 984-4611
EECS office -msg- 984-4482
home        (408) 353-3857

\←S\→L
Enc: 
  The Educational Issues
  The Bankruptcy of Basic
  Functional Flesh Functional Bones
  Duck Soup
  Faculty Workshop Proposal
  LISP on Personal Machines
  Resume







\C\F2 The Educational Issues
\C Computer Literacy: Technical and Humanistic\F1
\J
I don't separate these issues. One cannot  view the computer
as a strictly technological tool, independent of its potential impact
on society; neither can one discuss social issues related to computation
without having a grasp of the fundamental principles. There are too
many mindless discussions on both sides of the "street".

Futhermore, one cannot expect substantial inroads in
educational applications unless the appropriate tools are available;
and one cannot expect intelligent applications of LISP-like tools without
developing the  talent  to use them effectively.
Unfortunately, this approach requires an enlightened combination of
educational/publishing and hardware/systems organizations. As yet I have
not encountered such paragons of intellect. Computing concerns express
interest in my LISP expertise, and publishers express interest in the
text book segments; but one without the other is only half the battle.

Alas, the  educational woes are also substantial.
I am teaching a novel  course in Computer Science (EECS129 - The Art
of Computer Science)  
at the University of Santa Clara. Though student reaction has been quite
good, there has been some faculty irritation, even to the point of a
suggestion to get the class cancelled after the fact.
This course is an extended  version of the marathon tutorial I gave at  the
Computer Faire; it's chock full of technical substance, but my notion that
this material could be understood by non-majors troubles some, apparently.

Several beliefs drove my design of this course:

(1) That the fundamental ideas of computation are quite simple, being well developed
in mathematical logic in the decade of the 30's.

(2) As a corollary, I believe that the majority  of computing developments 
in the period between 1940
and 1980 represent technological, not intellectual advances. We now have
hardware to realize these early results --a distinct advantage, but the
view of computing that has grown up distorts and limits one's perception
of the computational phenomenon; assembler, Fortran, Basic, Pascal, and ADA
all deal with a superficial manifestation. Unfortunately, the traditional
 approach to computing begins with the symptoms and not the cause, and
one seldom sees the principles.

(3) Next, there is an emerging awareness that the traditional conception
of computers is  outmoded. Two new languages, LOGO and Smalltalk, are  becoming
publicly available. These "personal computing" languages bring the computing
medium into the realm of  human-oriented computing. These two languages
(developed and MIT and Xerox, respectively) are (a) "fun to use" and (b) directly
related to those early mathematical results.

(4) The gap between the technical and the humanistic view of the
world is pointedly illustrated by their shared misconception of computation:
one seeing the physical phemonenon as an end in itself, the other seeing the
same phenomenon as an impersonal tool, devoid of intellectual substance.
The engineer is at least as misguided as the  humanist;
a classical case of the Sufi elephant. The approach in 129 shows that
\F3both\F1 views are grossly inaccurate.
\.


\C\F2The Course Structure\F1

\J
EECS 129  addresses these issues in the following ways:

(1) Introduce the "fun" aspects of LOGO.  This language runs on Apple
computers, and  astounds the engineer as well
as the humanities person. As Pirsig remarks in \F3Zen and the Art of Motorcycle
Maintenance\F1, "Assembly of Japanese bicycle require [sic]  great  peace of
mind". The first week   of 129 was designed to establish the "peace of mind".

(2) Next we introduce the principles. Both "sides of the street" are at an equal 
disadvantage; the principles are no more complex than high school algebra.
Indeed, we began that way and  developed computational ideas from that simple base.
At the end of the second week they have seen some very powerful 
computing ideas, much more powerful than they yet realize.

(3) Build from those principles to the emerging view of computing as exemplified
by LOGO and Smalltalk.

(4) Explicate the mechanisms that make these very powerful languages look
so simple. This can be driven toward the theoretical end of the spectrum
in mathematical logic, or can be driven toward Computer Science practice
in LISP,  the ancestor of both LOGO and Smalltalk.

(5) Relate these ideas to cultural changes. In particular, I find Spengler's
study of mathematics and culture particularly seductive. I believe there
is an emerging "Computer culture" in a very 
non-trivial sense. I have conjectures that fit the
Spengler model from the mathematical perspective; a further issue is the
connection of these ideas with other  phenomena in our society. 
Regardless of the
outcome of this hypothesizing, the effort in examining cultural interrelationships
will be worthwhile.
This "non-technical" component of EECS129 is, to me, at
least as important as the elaboration of the technical components.

(6) As a practical undercurrent, I designed this course to  utilize an
interactive programming laboratory. Part of the computing phenomenon
involves using the machine. This  lab is not to indoctrinate students
with the rituals of hardware, software and typing; however, the newer kinds of
computing tools involve  a "culture shock" for those who are used to
the limited scope that traditional languages and machines  supply.
Indeed, a discussion of programming style --Pascal/ADA:Discipline versus 
LISP/LOGO/Smalltalk:Interaction is almost a religious issue.
It is important that students understand these issues.
\.




\F2\CThe Difficulty\F1
\J
The difficulty with the course is the lack of machines. Currently, we are
borrowing two Apples from San Jose State on a part-time basis; I have
bought some hardware to upgrade them to run LOGO. I have also borrowed money
to buy one TRS-80 Model II that I will be using shortly to introduce LISP.
Indeed, an unfortunate situation; and one I brought on myself. Last
September, when I proposed this course, I had strong encouragment from
Texas Instruments that they would supply hardware, because they had "great
interest in education". Unfortunately, their "interest" was  diverted.
In the intervening months, at least a half-dozen computer companies and
four book publishers have taken much of my time to explain the proposal
and, in the final analysis, said it's a great idea but either "call us
when it's finished" or "all our funds and machines are tied up".
This week  I ran out of potential sources, and physical and emotional energy 
(I ran out of money long ago).

I believe it is \F3very\F1 important that this experiment succeed; it
represents a fresh breath in computing. If it fails it will be all too
easy for all concerned parties to fall back into their previous patterns.
The change \F3will\F1 have to come sometime in the future   but
in the interim the students and the University will have lost valuable
time.
\.


\C\F2A Solution\F1
\J
Fortunately Dr.'s Mike Burke and Lois Flynne are anxious to continue
what I've started, so I soon expect to see versions of this course at San
Jose State and San Francisco State. At least one book will come out of this
and I  hope to move on the middle areas of education --high-school
and college-- much like LOGO and "Mindstorms" will do in the
primary areas.  I am also considering
the possibilities for a pair of videotaped
courses, one in the spirit of EECS129 and the other
specifically aimed at  the   AI applications
 audience.
Of course, such  moves take  substantial capital
and faith.  I feel it will pay off both intellectually and financially:
intellectually, because the alternative is a generation of technically
trained  but poorly educated computer engineers (see \F3The Bankruptcy of 
Basis\F1); financially, because the educational market, particularly in
computer-related fields, is immense.

Two hardware solutions seem viable. First,
one could utilize local networks of personal machines, 
as discussed in my letter. The alternative, which may be more cost-effective
is a single-box computer system. The shell contains the sharable hardware
(power supplies, disks, and printers)
and provisions for installing from one to a dozen single-board computers.
These single board computers would be self-contained, say: Z-80A/B, 64K user
ram, at least 4K video ram (bank-switch addressable), and provisions to
transmit video output, receive keyboard input, and communicate with the
shared devices.
The case would allow  one to plug in keyboards and monitors, adding
equipment as needed or affordable.  The video output could be switchable in the
sense that several monitors could be "mapped" to the same screen, allowing
flexible interaction between users and/or instructors. Of course, one
could apply this scheme to keyboard input as well. In this scheme  the intelligence
is in the box, piping only video or key-stokes to and from the outside world;
this way one can update the processing power of a single card without
investment in new monitors or keyboards. The flexibility is very attractive.
\.


\C\F2 The Educational Issues
\C Technical Education\F1
\J
This area is an outgrowth and elaboration of the basic principles
laid down in the previous courses. Given the fundamentals, we can build
applications of the LISP ideas:  artificial intelligence, complex system
design in general, applications of intelligent systems. All of this
material applies the principles that the basic course develops.

Of course, such applications require more computing power than that
available on the less expensive systems; but the theoretical and
practical knowledge scales up naturally. The interactive programming
style migrates, as does the knowledge of how to use the tools in a
craftsman-like fashion.

Here we would expect to utilize the high-end 68000/16032 systems.
\.


\C\F2Direct Technical Contributions\F1
\J
We have already mentioned the systems aspects. Their potential utilization
is now a question. As Business Week, Wall Street Journal, and other publications
are reporting, AI is moving out of the research environment into commercial
applications. A critical component of the technology transfer is the availability
of inexpensive LISP systems. As yet this market is open; incisive action can
establish TLC in this area. ARPA, the Navy, and other government agencies
have been waiting for about two years now for software to be ported to 
appropriate new machines.
\.


\C\F2Future Trends\F1
\J
Times are changing: LOGO has begun to move on the low-end LISP-like market;
LMI and Symbolics are  embattled in the high-end arena. The middle
ground is still open, and it is that area that I want for TLC, both
educationally and technically.

Smalltalk-80, will also have an across-the-board impact in proliferating
LISP-like ideas. It is expected that five manufacturers will  announce their
implementations sometime this year, with some announcements expected
at NCC. I believe that Smalltalk's influence will have a very positive effect on
LISP's popularity, for the ideas that support Smalltalk (functional objects)
are easily grasped when seen in the rarified light of LISP even though
they are be-clouded in Smalltalk's syntax: LISP is to Smalltalk as
mathematics is to engineering. The combined power of LISP, LOGO, and Smalltalk
should have significant effect on the world's perception of  computational ideas.

More directly related to LISP and TLC is the application of the tools and
the minds that the process will develop. This is the business of applied 
AI. That period is sufficiently distant from the realities of TLC's current
position that I'll refrain from hypothesizing.
\.


\C\F2Summary\F1
\J
A broad plan to change the world; arrogant, perhaps, but the alternative
of a populace weaned on Basic and word-processing is not at all attractive.

I  believe we can do it. I'm reminded of a short story I read a long time ago.
It seems that in the middle of this desolate plain, several major trade routes
intersected at this giant rock. In the nooks and crannies of this rock the
 beggars and wayfarers  huddled to get out of the wind and exchange comments
about how poor was their lot in life. One day a band of men with a team of
oxen came to the rock, drove away the beggars, and pulled the rock from the
ground. They  proceeded to dig a pit and extract from its bottom untold
riches, which they carried off. The beggars then went back to their rock.
I think it's time we moved the rock.