Reprinted from SCIENCE BOOK REVIEWS 17 December 1965 Computers and the Life Sciences Any writing on computers faces the hazard of instant obsolescence. The “second generation” computer is now a rapidly aging six-year old. In the time needed to obtain some perspective and win even a glimpse of consensus on the career prospects of this youngster, he is already a technological misfit. “7090” could evoke pure awe from the freshman, serve as a grudging partner of the sophomore and a workhorse to the junior, and then graduate as an antique like the model-T or the DC-7. Professors and monographers are less flexible and more sentimental than their students, for they find it harder to shed the milieu of their own indoctri- nation. However, we certainly need their books and monographs even if the gestation time conforms more closely to human than to computer- mechanical time scales. Four recently published treatises are reviewed here— Use of Computers in Biology and Medi- cine (McGraw-Hill, New York, 1965. 989 pp., $29.50), by Robert S. Ledley; Computers and the Life Sciences (Co- lumbia University Press, New York, 1965, 352 pp. $12.50), by Theodor D. Sterling and Seymour V. Pollack; Com- puters in Biomedical Research, vols. 1 and 2 (Academic Press, New York, 1965; vol. 1, 584 pp., $20; vol. 2, 363 pp., $14), edited by Ralph W. Stacy and Bruce Waxman; and Mathematics and Computer Science in Biology and Medicine (Her Majesty’s Stationery Of- fice, London, 1965. 327 pp. £3), edited by Harold Himsworth and George Godber. Use of Computers in Biology and Medicine and Computers in the Life Sciences are didactic efforts which should benefit from unity of author- ship, That advantage is possibly out- weighed by their inevitably longer ges- tational delay and their preoccupation with an implicit hardware set that has diminishing relevance. This is a special Joshua Lederberg is professor of genetics in the School of Medicine at Stanford University, Stan- ford, California. 1576 Copyright © 1965 by the American A Joshua Lederberg difficulty in Ledley’s volume, most of which must have been written at least 5 years ago. In fact, the potentially most interesting sections have lost more than they have gained in their revision from his previous works. His predilec- tion for a three-address machine may have some design basis, but is increas- ingly at odds with the actual trends in the field. The same engineering bias is apparent in the rather confusing and inefficient treatment of the program- ming languages which are the main- stream of computer science. It will be enough to note that “FORTRAN,” “list- processing,” and “assembly language” are absent from the index and to quote from page 83: “The fewer addresses that appear in the instruction format, the more complicated the coding be- comes and the greater the number of instructions required in a code. Thus automatic-programming aids become especially important in one-address in- struction formats.” As much as I admire Ledley’s exposi- tion of hardware design, I must deplore this approach to teaching about com- puters, even for teaching engineers, and much more as an approach for teach- ing the use of computers for research applications. A large part of his book is a textbook for an undergraduate course in mathematical analysis. This is not directly related to computation, but may reflect the mathematical so- phistication needed for analog compu- tation and for some of the simulations of biological systems which make up the rest of the work. On the other hand, there are many valuable nuggets strewn over the landscape, and for these nuggets alone I would recom- mend his book to any investigator. Computers and the Life Sciences is a more coherent presentation of an out- look on the uses of computers as these uses might be seen at a university served by a medium-sized computing center. In view of current trends, the univer- sity system outlook deserves even more emphasis than it begins to receive here. for the Ad A good general purpose introduction to computational techniques is hard to write, partly because of the rapid rate at which it becomes obsolescent, partly because of the enormous dispersion of the audience’s interest and preparation. Meanwhile, as a work addressed to ex- perimental biologists, Sterling and Pol- lack’s volume has few if any competi- tors. Although it stresses the problems of closest concern to the authors, and especially the planning of radiation therapy, the volume is sensitive to the familiar languages and work habits of biomedical research generally, and emphasizes analog inputs and statistical processing. Computers in Biomedical Research, volumes 1 and 2, which calls on the specialized experience of a large and illustrious group of contributors, is cer- tainly the most timely and broadly interesting of the works reviewed here. Most of the authors do not presuppose a detailed background in computer techniques for their readers, and many of the chapters could be read as an introduction to computer applications by specialists in the various fields of biological research. These two volumes edited by Stacy and Waxman already cover a very wide range of topics and points of view. We may hope that they initiate a badly needed, continuing series of state-of-the-art reviews. The chapters are in the main very well writ- ten, to the point, and related to the full range of significant contemporary ap- plications: analog to digital conversion and data handling, especially for neuro- and myophysiology; simulations of com- plex behavior; analyses of chemical reaction chains; and x-ray analysis; but they provide only some glimpses of computer-controlled hospital manage- ment, and even less of automated li- brary processes and laboratory routine and of programmed teaching systems. These volumes lay strong emphasis on demonstrated accomplishments, an al- most unique virtue in the literature of this fast-moving field. In consequence, however, the reader may not sense the significance of the new systems that are just coming into the field. Versatility of access to the system rather than brute- force computing power is the hallmark of this era, and of course quite indis- pensable for the general use of the computer as a workaday tool by every scientist in his own taboratory: com- putational service as a public utility. In some ways Mathematics and Com- puter Science in Biology and Medicine comes closest in flavor to this anticipa- SCIENCE, VOL. 150 tion of the future which is rapidly be- coming the present. As the proceedings of a conference with recorded para- phrases of the discussion, it condenses a great deal of intelligent thought on problems and uses of computers. The expositions on the logical problems of classification, on data systems for hos- pitals and large-scale vital statistics, and the commentaries on the extraction of significant wave forms make this another indispensable volume on the reading list of any quantitatively minded biologist, even apart from its use in orientation toward computer applica- tions, The contributions thus give an excel- lent perspective on the uses of compu- tation for biomedical research. For many biomedical researchers today, the computer’s most important uses are perhaps in signal analysis and in x-ray crystallography, and all of these works pay considerable attention to analog processing and digital numerical anal- ysis, For these applications, the digital machine is simulating an analog device which, except for inflexibility in repro- gramming, could weil outperform the digital machine. Stress on such simula- tions might obscure the role of the digital computer as an automaton whose fundamental processes are in fact not numerical but logical. Indeed, great in- genuity has had to be applied to achieve the realization of arithmetic operations as complexes of the logical steps inher- ent in the computer. This concept is of course underlined as an aspect of engi- neering design, and it is fundamental to Newell and Simon’s chapter “Pro- grams as theories of higher mental processes” (in Computers in Biomedi- cal Research, vol. 2). Thus, the utility of the computing machine as a numer- ical analyst may eventually awaken the biologist’s attention to an event of phylogenetic rather than historical sig- nificance: the emergence of a symbol- manipulating organism which is now well launched on a distinctive evolu- tionary pathway. If many a human being congratulates himself that his power to turn off the machine is still the decisive fact of this symbiosis, he may be rudely disillusioned—say by a spell in prison—should he violently attack an economic value on which many hundreds of his colleagues de- pend. The phylum Automata is evolving rapidly—doubling its general capability at intervals of about 15 months—ac- cording to a new set of rules that we dimly understand. Even more conse- 17 DECEMBER 1965 quential than the hardware are the pro- gramming systems and languages which have by no means exhausted the capa- bilities of last year’s hardware. With his concern for the forces that underlie organic evolution, the biologist has a special responsibility to understand the phylogeny of automata as a major com- ponent of the orthogenesis of the plan- et. These works are not directed to such ethereal issues, but a by-product of their educational message to biol- ogists is an appreciation of the micro- scopic anatomy and system physiology which must underlie the ecology of these new organisms.