College and Research Libraries 152 I College & Research Libraries • March 1971 sibly suited for information retrieval (with reservation). Conspicuously lacking in the discussion is any mention of SNOBOL or PL-1. This discussion is followed by men- tion of the primary types of computer file structures, e.g., sequential, list, and tree files. A chapter is devoted to information re- trieval systems. The chapter covers most of the important work done in the field: co- ordinate indexing, thesauri, SDI systems, KWIC indexing, automatic classification ex- periments, clustering, and so forth. Each topic is mentioned and briefly discussed. The last third of the book is devoted to examples of IBM 1620 computer programs written for a library system. The data input program, the sort program, and the search program are presented in great detail, in- cluding sample output and ten pages of flow chart. These programs essentially con- sist of the author's efforts at an experimen- tal searching program to retrieve author names by variant spellings. Although the search program is of some interest, such long and detailed examples appear out of place in such a short book. It is material better suited for a journal article. My major objection to the book is not that it is superficial (since this can be turned into an asset in a general introduc- tion to a subject), but rather that the book is fundamentally misleading for the unin- formed reader (the only type of reader for which the book is appropriate). It gives the impression that to develop an automated library system all one need do is buy a com- puter and hire a programmer. Even in the final chapter, a discussion of the "philoso- phy of a machine-based system," the only difficulty mentioned is that of automating fact retrieval. Although we all seek simple and straightforward solutions to problems of library automation and information re- trieval, it seems a disservice to those enter- ing the field to give the impression of sim- plicity.-Victor Rosenberg, University of California, Berkeley. Systems Analysis of a University Library; Final Report on a Research Project. By Michael Buckland, et al. (University of Lancaster Library Occasional Papers, no. 4). Bailrigg, Lancaster, University of Lancaster Library, 1970. 57p. The application of operations-research techniques to libraries is a recent phenome- non, and most library administrators prob- ably shudder at the sight of equations pur- porting to represent the behavior of some part of their library system. Some people shudder at the sight of any equation, but many are also concerned about what seems to them the simple-minded nature of the idea that a library process can be represent- ed by equations at all. The present report presents two kinds of answers to concerns such as the latter class of readers express. On the one hand, the authors candidly ad- mit that their work is fragmentary and ex- ploratory, and can hardly be anything more, given our ignorance of our users' be- havior. (What on earth do faculty members do with · those books that they keep for six months?) But they also point out that it is studies such as the present one that help to point out just what it is we need to know before more advanced work can be under- taken. The second answer that the authors can make to persons skeptical about the value of operations research is to point to prac- tical results. Their studies of the effect of various loan policies upon the availability of books gave them sufficient evidence to induce the University of Lancaster to ac- cept important changes in loan policies- the evidence being that the new loan policy would, by all indications, appreciably im- prove some aspects of the library's service to its users. Another study showed that at a small increase in annual cost, the library could establish a bindery (it had been using commercial binders) and thereby ef- fect other improvements in service. This report then reflects both the short- comings and some of the potential values of operations research as applied to li- braries. Its title may be misleading for it by no means reports the results of a systematic analysis of a complete library system. It is rather a report on several studies of various facets of a library system. There is, for ex- ample; ·a section on the frequently discussed and a'nalyzed problem of acquisition and retention policies, with particular attention to the implications of relying upon interli- brary loan for access to titles either not ac- quired or discarded after being held for a period. Another section develops mathe- matical models of acquisitions policies for libraries that are part of a hierarchy of li- braries consisting of personal libraries, de- partmental libraries, university libraries, and a national library. In another section, the authors speculate upon the pervasive- ness in librarianship of negative exponential distributions similar to the famous Zip£ dis- tribution. They argue persuasively in just a few paragraphs that it is important from both the theoretical and practical points of view to find out to what extent their specu- lations are correct. This report will be useful to library sys- tems analysts and operations researchers. It would also be a good document for read- ing by library administrators who would like to get an idea of the kinds of things that are going on in library operations re- search. (Those whose forte is not mathe- matics should skip over the detailed mathe- matical expositions. All significant points are also made in prose.) Finally, this is the kind of document that a university librarian might well want to share with some of the operations research specialists on his cam- pus. Collaboration between librarians and operations researchers outside the library has already borne fruit at several institu- tions. Any library ordering this report should probably also order the Library's Occasion- al Papers nos. 1, 2, and 3, since some of the material in this no. 4 is merely a condensa- tion of work reported more fully in the three earlier papers. The reader should be warned that there are numerous typograph- ical .errors in the equations and in the dis- cussions of them.-Kelley L. Cartwright, School of Library Service, University of California, Los Angeles. The Subject Approach to Information. By A. C. Foskett. Hamden, Conn.: Archon, 1969. 310p. $7.50. It is important to view this volume in the context in which it was written. In his preface the author states that "it is my hope that this work will help to fill a need Recent Publications I 153 which exists at present for a textbook suit- able for that part of the Library Associa- tion part I paper 3 syllabus which deals with the subject approach." He then goes on to say that he has "tried in this text to emphasize the similarities as well as the differences between the different methods of retrieving information, and to show the kind of situation where each is likely to prove of value." To accomplish this in modem context he has "concentrated ev- erywhere on the modem approach, includ- ing terminology .... " The volume was examined in terms of both form and content. There are minimal variations in typeface which make it diffi- cult to separate the examples from the text. The arrangement is logical for the sub- ject it covers, going from· the general to the specific. However, in arranging the text where it concerns particular schemes, a general discussion is used with examples before the general layout of the particular classification scheme under discussion is given. The index is thorough and useful. Nineteen percent of the indexed items re- fer to bibliographic data. Since the con- tent is generally aimed at a British audi- ence, the systems in use in Great Britain and India are more thoroughly covered than those in use in the United States. Whether the chapters are arranged in the order of importance to the author, or whether, by so arranging them he has eliminated some redundancies, it is inter- esting to note that the number of pages de- voted to each system decreases (except for Colon): DC 23 UDC 15 Bliss 12 Colon 14 LC 9 LCSubjHd 5 Sears 3 There are some minor problems in lan- guage. For example, when Foskett discusses homographs (p.40), he lists PITCH PITCH PITCH (Bitumen) (Music) (Football), in our context we would use (Baseball) ;