m a y be used procedures. as arguments of other (or the same) Some users of C o m m o n Lisp or SCHEME m a y d i s agree about t h e i r language's inability to use procedural a r guments. For instance, in Winston and Horn's w e l l - k n o w n book, LISP, 2nd Edition, Chapter 4 has a subheading titled, "Recursion A l l o w s Procedures "1"o Use Themselves." They then proceed to give an example of a procedure using i t self as a part of its recursive definition. This w o u l d seem to be a serious lapse of scholarship on the author's part. Whether this chauvanism for his chosen language should be an i m p o r t a n t f a c t o r in deciding to use this o t h e r w i s e - e x c e l l e n t b o o k is up to the i n dividual. It is possible that the w a y T uses procedural a r g u m e n t s is different f r o m that described in Winston and Horn, and it m a y have s o m e superior quality unknown in SCHEME or C o m m o n Lisp. Aside f r o m the procedural a r g u m e n t problem, the T P r o g r a m m i n g Language is highly r e c o m m e n d e d for the beginning Lisp p r o g r a m m e r . Its organization, clarity, and generous use of examples set it apart f r o m t e x t b o o k s this w r i t e r has p r e v i o u s l y e n c o u n t e r e d in the effort t o b e c o m e Lisp-literate. Speaking philosophically, I personally w o u l d prefer to see AI t a u g h t as a course in advanced p r o g r a m m i n g t e c h niques for clever programs. It really is t o o easy to c o m e up with c o u n t e r examples to the brain as a m o d e l for the mind: All one has to do is ask h o w m a n y pixels are sent to the brain by each eye and h o w m a n y bits w o u l d be needed to represent even even one brain i m a g e in a c o m puter. The arithmetic for such a g e d a n k e n - e x p e r i m e n t i m m e d i a t e l y shows that, even with 50 billion or so neurons, v e r y few images could be retained for recall if the brain w e r e d e v o t e d solely to the storage and recall of images. AI researchers w o u l d do well to protect t h e m s e l v e s against future ridicule by couching their p r o g r a m m i n g e x amples in metaphors less suggestive of an uncritical belief in neurons and neuron nets. The Elements of Artificial Intelligence An Introduction Using LISP by Steven L. Tanimoto Computer Science Press Reviewed by: Kenneth R. Lee Rider College Lawrenceville, New Jersey A QUICK REVIEW OF THE ELEMENTS OF ARTIFICIAL INTELLIGENCE: An Introduction Using Lisp Steven L. Tanimoto Computer Science Press, 1987 Victor Schneider C o m p u t e r Science Department Framingham State College Framingham, MA 01701 This is a really fine b o o k for a first course in AI. Its examples are w r i t t e n in a subset of C o m m o n LISP c o m patible with o t h e r LISP dialects, so that t h e y can easily be rewritten to run on say, CDC ALisp, or p u b l i c - d o m a i n , DLisp, or s o m e v e n d o r - s u p p l i e d system, like Translisp. (I have his version of Eliza running on XLisp and Translisp with m i n o r changes. The "Doctor" p r o g r a m in the earlier edition of Winston's AI t e x t b o o k will also run in place of Tanimoto's "Shrink," with only m i n o r changes.) While the version of LISP used in the T a n i m o t o book is not nearly as clever and baroque as those of the o t h e r currently available LISP i n t r o d u c t o r y books, there is enough intricacy to challenge students w i t h o u t forcing t h e m to learn advanced features of C o m m o n LISP unlikely to add to their general understanding of what LISP is all about. While the book title does not m e n t i o n it, the book also features a discussion of Prolog, including a Prolog in LISP i n t e r p r e t o r and explanations with i m p l e m e n t a t i o n s of the Prolog unification algorithms. Wi{h a solid Prolog book as companion, such as I. 8ratko's Prolog and AI text, this book gives a sound p r o g r a m m i n g introduction to the major areas of AI. The b o o k tends to favor "scientific" AI application, such as image processing and comprehension, with s o m e w h a t less emphasis on expert systems, a l t h o u g h expert systems are covered. The major flaw that I see in this book, as well as in o t h e r AI texts is in its reliance on an underlying biological m o d e l that is u n p r o v e n and m a y be incorrect in the last analysis. This reliance appears in the uncritical acceptance of the brain as an adequate model for the human mind, and neural nets as adequate m o d e l s for learning and p r o b l e m - s o l v i n g . The ligence: Elements of Artificial Intelan I n t r o d u c t i o n using L I S P is in- tended (as stated in the preface) for use as the t e x t for an i n t r o d u c t o r y course in artificial intelligence. The author's approach to the subject is stated on the dust jacket (and in the preface, although not as concisely). "The rapidly expanding subject of Artificial Intelligence requires professionals w h o have a firm grasp of both its scientific principles and its i m p l e m e n t a t i o n techniques. Without the principles, the p r a c t i t i o n e r flounders w h e n e v e r he reaches the limits of his tools. Without i m p l e m e n t a t i o n experience, one has very limited intuition about w h a t is feasible and h o w a new project should be organized." The author's goal is, therefore, to elucidate the principles of artificial intelligence and, t h r o u g h a series of p r o g r a m m i n g examples and exercises, to p r o v i d e practice in applying those principles in producing w o r k i n g systems. Reading the text, it is very clear that the a u t h o r is very k n o w l e d g e a b l e a b o u t this area and has had a lot of e x perience to draw on. The t e x t covers the entire breadth of the field and presents classic problems and results in each area t h r o u g h the p r o g r a m m i n g examples. Despite the author's k n o w l e d g e and experience, the breadth of coverage, and, I w o u l d add, an excellent order of presentation of the areas of artificial intelligence, this book does not, however, in m y estimation, succeed in f u l filling its stated goal. To a m i n o r degree this is due to • mechanical problems with the book (e.g., the w o r d "belief" is used on page 2 w h e r e "believe" was intended, the first part of section 4.5.2 is found repeated on pages 95 and 96). These problems seem to indicate t h a t the a u t h o r a l lowed the use of a spell-checking p r o g r a m to substitute for a careful p r o o f - r e a d i n g of the text. More importantly, however, the book suffers f r o m a confusion of the author's t w o purposes. It is often unclear in particular sections w h e t h e r the author's intention is to present a principle and its implications or to demonstrate the i m p l e m e n t a t i o n of a principle. Thus the understanding of the principle b e comes confused and the connection of principle to i m plementation becomes tenuous. SIGART Newsletter, January 1988, Number 103 Page 20

/lp/association-for-computing-machinery/book-review-a-quick-review-of-the-elements-of-artificial-intelligence-5Qm17FzNkt