Essential systems analysis

i n reference form with each chapter. The text is not sufficient i n itself to give the reader the knowledge depth necessary to design a complete robot system but it does identify the essential tasks and leads the reader to apoint where subsequenttasksare well defined and thoroughly referenced. Critchlow’s introduction to Robotics is an essential addition to the practicing engineer’s robotics and computer-control engineering library. Reprinted from / E E E I. Robotics and Automation, vol. RA-1, no. 4, p. 215, Dec. 1985.

Essential Systems Analysis-Stephen M. McMenamin and John F. Palmer. (New York, NY: Yourdon Press,1934,392 pp., $29.95.) Reviewed by Donald R. Chand, ComputerInformation Systems Department, Bentley College, Waltham, MA. Before a quality system can be developed, its requirements must bespecified. The structured techniques of the 197Os, which markedly impacted our ability to design quality systems, do not fully address the problems of developing the true requirements of a large system. It i s heartening to see a book that does. A primary contribution of McMenamin and Palmer is the way i n which they focus on the problem. They assert that developing the true requirements entails discovering the essence of the system. This essence is defined as the logical requirements of the system, assuming thataperfectimplementation technology is available. Derivingthe essence of systems and making it the basis ofa systemsanalysis methodology is the central theme of Essential Systems Analysis. The 31 chapters are grouped into eight parts. The first introduces the fundamental concepts underlying the essential systems analysis methodology. The authors postulate that the essence of a system can be captured i n terms of essential memory and essential activities (classified as fundamental, custodial, and compound essential). The concept of perfect technology i s explained and the impact of imperfecttechnology on any incarnation of the essence is discussed. In the second part, the tools and strategies for modeling essential activities and essential memory are presented. Event partitioning i s recommended as the strategy for identifying essential activities, and object partitioning is suggested as the means of structuring essential memory. Once events are specified, essential activities are modeledin terms of the planned responses and results of the system and essential memory accesses.Essential memory, on the other hand, is modeled i n terms of objects and the intraobject relationships. Having introduced the fundamental concepts of essential activities andmemory,their representation, and basicstrategies for modelingthem,the authors deal with the actual problem of deriving systemessence. In the next four parts, techniques for derivingthe essence of an existing systemarediscussed.The seventh part outlines a strategy for deriving the essence of a new system.Thelast part contains recommendations for managingthe tasks involved in deriving the essence of large systems. This book offers newinsight on the goals of systems analysis and provides many good ideas for deriving the requirements of large systems. Since the book is packed with new terms and assumes a basic knowledge of the tools and techniques used in structured analysis, few practitioners could assimilate the essentialsystems analysis techniques by reading the book alone. I also feel that the book would not be a good text in a systems analysis course, since the authors give no end-of-chapter exercises or review questions. Also, the authors appear to have writtena monograph documentingtheir understanding of the subject. Instead of simply defining the concepts, illustrating them with examples, and then showing their application to a case, McMenamin and Palmer defend and justify their ideas of essentialsystemsanalysis as better than the analysis techniques oriented toward data flow structuring, Despite these shortcomings, the book makes a definite contribution to systems analysis literature, and anyone teaching a course on the subject should read it.

Book Alert

The following descriptionso f recent bookswere prepared by the staff o f the EngineeringSocieties Library, 345 East 47th Street, New York, NY 10017-2394. These books are available in the Library for loan or reference use. The prospective buyer should contact the listed publishersor his local technical book store.

Optical Bidability: Controlling Light with Light-Hyatt M. Cibbs. Orlando, FL: Academic Press,1985,471 pp., bound, $54.50,ISBN 0-12-281 940-3.) This is a research bookintended for new entrants and active workers i n the field of optical bistability. The treatment interprets opticalbistabilitybroadlytoincludeallofthe steady-state and transient characteristics of nonlinear optical systems which exhibit bistabilityunder some operating conditions. It is restrictive i n placing the emphasis on passive (non-laser) systems which exhibit reversible bistability with input intensity as the hysteresis variable. The book i s motivated by the desire to summarize the beauty of the physics and to describe the potential applications of such systems for nonlinear optical signal processing.

Theory of Microwave Remote Sensing-Leung Tsang et a/. (New York, NY: 1985, 613 pp., bound, $44.90, ISBN 0-471-88860-5.) This book studies the active and passive microwave remote sensing of earth terrains such as snow, ice, soil moisture, vegetation cover, forestry, cloud, and rainfall. Recent developments in the theoretical models that have been proven useful in the interpretation of the experimental data are presented. Basic principles of electromagnetic wave propagation and scattering are used to derivetheequations that govern microwave remote sensing.The modeling, derivations, and resultsare presented i n a detailed manner.All relevant topics in the modeling of earth terrain are covered, providing a valuable reference for researchers and engineers and a solid background for further research.

Integrated Services DigitalNetworks-Anthony M. Rutkowski. (Dedham, MA: Artech House,1985, 324 pp., bound, price not given, ISBN 0-89006-146-7.) This book portrays the development of ISDN from an esoteric concept and dream of advanced information network planners to a set of worldwide standards adopted by CCITT, and finally to the initial equipment and systems of today. The book presents and explains technical and policy material emanating from international and domestic forums. It also contains proceedings, and actions by the FCC and other federal agencies.

Bandit Problems; Sequential Allocation of Experiments-Donald A. Berry and Bert Fristedt. (New York, NY: Chapman & Hall, 1985, 275 pp., bound $25.00, ISBN 0-412-24810-7.)

Reprinted from /€€E Software, vol. 2, no.4, pp. 110-1 11, July 1985.

This book provides a comprehensive and general account of banditproblems starting from first principles. Both discrete and continuoustime are considered. Many examplesare given that both convey the spirit of bandit problems and demonstrate techniques for solving them. Important results reported in the literature are reviewed;butinaddition, many new results are given-for example, the converse of the famous Cittins-Jones result is proved. Results not covered within the text are given within an annotated bib liography of some 200 items.

PROCEEDINGS OF THE IEEE, VOL. 74, NO. 10, OCTOBER 1986

1471