COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	\\M1BASL30\M2BASB30\M3NGR25\M4NGR20\F2\CSTANFORD UNIVERSITY
C00011 ENDMK
C⊗;
\\M1BASL30;\M2BASB30;\M3NGR25;\M4NGR20;\F2\CSTANFORD UNIVERSITY
\F3\CSTANFORD, CALIFORNIA 94305
\F4COMPUTER SCIENCE DEPARTMENT\←L\-R\/'7;\+R\→.\→S   Telephone:
\←S\→.415-497-4971
\F1\CDec 4, 1974



Blake E. Vance, Assistant to the President
American Elsevier Publishing Co., Inc.
52 Vanderbilt Ave.
New York, N.Y. 10017


Dear Mr. Vance:

\JFinally, here it is. After talking with you, I made copies to send
and began constructing a letter outlining the revisions I wished to make.
It turned out that the explanation of the revisions was more complicated
that just making the revisions.  So I (sigh) made the revisions.

Though the revisions have been made, the additions must wait until
January, but the current manuscript with the "apology"
should express my intentions.
These additions have been worked out in detail during my lectures at
UC Santa Cruz, so I forsee no difficulties in incorporating them.

Let me give you a slight introduction to  the manuscript. It grew out
of  a course on data  structures I developed at  UCLA. I expanded and
revised the notes here at Stanford, and have taught variants  at the
UC Santa Cruz graduate  workshop the last  two years. I will  be helping
San  Jose State  set up their  graduate data  structures courses next
semester  and will  use  the  manuscript  there. 
I will be teaching their first data structures course in the
spring term. A  few  students  in
Stanford's CS206 course used it,  and students of Tony Hearn at Utah
are currently using it.
The reaction from rewievers, both students  and the research community
has bee quite favorable.


The manuscript will attempt to be self-contained because I think LISP
is the best way  to introduce prospective students to  the field. The
fewer preconceptions about  programming languages,  the better.  Thus
the book  goes  from very  basic  undergraduate material  to  current
research in semantics.  The material on λ-calculus and Scott's models
is  currently  very  rudimentary.   I  am  attempting  to  develop an
intuitive introduction to these areas.

Basically the material falls into  5 areas: 

	\F21.\F1 The mechanics  of the
language;  recursion,    functional  arguments.    Representation  of
algorithms in a  programming language, representation  of domains  as
data structures. 

	\F22.\F1 Evaluation; the importance  of interpretation and
its  relation  to  denotational models.    

	\F23.\F1 Implementation of the static structure of LISP and the
problems of machine orgainzation. 

	\F24.\F1 The dynamic structure of LISP and compilers for LISP;  LISP, being
a  very clean  way  to introduce  compilation.   

	\F25.\F1 Implications  for
language  design; given a clear  understanding of LISP,   what can be
done better? Most of this  comes from  my own ideas  on a LISP-like  language
with  user-defined data  structures  and a  semantics  which is  more
amenable to proofs and verifications.
Besides my own ideas, this section will incorporate other more well-known
"extensions" of LISP ideas. Namely the two innovations deriving from PLANNER:
pattern directed innvocation, and structured data bases. 

My writing style is informal, but I do not want to appear flippant or
be imprecise; either of these faults is deadly. Two reviewers objected
to the style; six made explicit positive comments.

Now,   the state of the manuscript.  Parts  (1),  (2),  and (3) are in
reasonable  shape; perhaps  85%  complete.
Part (4), on compilers, perhaps 65% completed. Part (5), though
absent from the manuscript is at least 50% completed. 

The reviews have been quite favorable,   and I am convinced that  the
manuscript  should be  published.    LISP  is the  most  most  poorly
understood programming language around, and it gets a little tiresome
to see people re-inventing McCarthy's ideas simply because there's no
decent documentation.  There's a succinct statement by Strachey which
I  think  represents position  and  the current  state of  affairs: "I
always worked on programming  languages because it seems to  me until
you  could  understand those,    you  couldn't understand  computers.
Understanding them doesn't really mean only being able to use them. A
lot of people  can use them  without understanding them."  This quote
appears at the beginning of the chapter on evaluation!


The current manuscript represents a entensive revision of the material,
incorporating many reviewer's suggestions. Though no one has had a 
chance to see this new product,
here are  some of the guinea pigs who read previous versions: D.  Bruce
Anderson, Dr. Friedrich von Henke, Dr. J. Strother Moore, Mike  J. Clancy,
 Hanan Sammet, Dr.  Anthony C. Hearn,
 Jorge  Morales.  Parts have been read by Dr. Lynn Quam, Dr. Michael Gordon.  
 In all,  15-20 "people" have read parts of the
manuscript; enumerable "students" have made comments.

Dr. Nilsson suggested that I mention possible reviewers.
Dr. J. S. Moore of the Xerox Research Labs, has reviewed a prior
version of the manuscript and made valuable suggestions on improving
the presentation. He is willing to review this later version.
\.
\←L\→S\←R\-L\/'2;\+L\→L

Yours sincerely,



John R. Allen
Research Associate
Computer Science Dept
Artificial Intelligence Labs

\←S\→L