Book Reviews 211 ACKNOWLEDGMENT The work reported in this paper was supported in part by a grant from the U.S. Office of Education, OEG-7-071140-4427. REFERENCES 1. Ruecking, Frederick H., Jr.: "Bibliographic Retrieval from Biblio- graphic Input; The Hypothesis and Construction of a Test," Journal of Library Automation, 1 (December 1968), 227-38. 2. Lipetz, Ben-Ami; Stangl, Peter: "User Clues in Initiating Searches in a Large Library Catalog," in American Societ'l for Information Science, Proceedings, 5. Annual Meeting, October 20-24, 1968, Columbus, Ohio, p. 137-139. BOOK REVIEWS Conceptual Design of an Automated National Library System, by Norman R. Meise. Metuchen, N.J.: Scarecrow Press, 1969. 234 pp. $5.00. This is a very confusing book. And it is too bad, because this reviewer kept feeling that the author, Norman Meise, had something to present. The trouble is that he does not communicate. This, I think, is the result of two things. First, the book reflects the naivete of engineers when they come to deal with what are basically social systems like libraries. This does not mean it can't be done, but such a task needs clarity and pur- pose, which this book does not have. The second springs from this failure. The masses of data, assumptions, and commentary in the book are poorly organized and interrelated. It is not enough to write strings of words; those strings must communicate and relate backward and forward in the text. Although never explicitly stated, the book evidently grew out of a study performed by the United Aircraft Corporate Systems Center in 1965-66 for the development and implementation of a Connecticut Library Re- search Center (see ERIC Document ED 0221512). The latest reference in the book is 1966. In a field, i.e. library networks, where a fair amount of work and discussion has taken place in the last three years (e.g. the EDUNET Conference in 1966), a book like this quickly loses its impact. The purpose of the book, according to the author, is "to show the feasi- bility of a system concept rather than provide a detailed engineering design." The system is "an automated national library system" using the State of Connecticut as a model. The author then adds (spoiling the whole introduction) : "If these functions (bibliographic searching, acquisition, cataloging, circulation) can be economically automated, the major prob- lems associated with our information explosion will be solved." As Anatole France once said: "It is in the ability to deceive oneself that the greatest talent is shown." 272 Journal of Library Automation Vol. 2/ 4 December, 1969 Basically the system is made up of three levels : local libraries, the re- gional center, and the "national library central." These are interconnected either by teleprinter, at 75 bits per second, or CRT consoles, as 1200 to 2400 bits per second. Mr. Meise develops extensive tables, using Connecti- cut as a model, for (a) estimated message traffic, real-time and batch; (b) allocation of communication traffic to segments of circuit route; (c ) cumulative communications traffic; (d) number of circuits required ver- sus circuit speed. He discusses bibliographic coupling ( 78-82), the Itek Memory Centered Processor, disc packs and file organization, ( 100-118, 162-179). I cite these tables and data (there are many more) merely to show the approach. At one point he talks about packages such as books, at another about papers. The whole system is based on statistics for which there is no discussion. Item: "the local library should satisfy a large per- centage of the user's needs ( 90-99% ) ; however, some portion of these needs ( 1-10%) should be obtained from other libraries to keep system costs within reasonable range" ( p.32). Where does "90-99%" come from? How do we know that this level will "keep system costs within reasonable range"? Item: "The State of Connecticut is about the right size for a regional center from the point of view of expected user load" ( p.118). Whose hat did he pull this one out of? There is no discussion of right size, nor really any of what "size" means -population? geographic area? cultural makeup? One suggested region (Arizona, Nevada, Utah) has about the same population as Connecticut, but is 62 times the size. Cer- tainly the communications costs are entirely different and the two regions are not comparable. Figures suddenly appear in the text, e.g. 9,610,000 vols. (p.98) and others, and the reader does not know where they came from. They may be right. They may even have been discussed somewhere in the text, but on page 98 one does not remember. And the index is of no value: two pages, hastily organized. This is all too bad, because Mr. Meise evidently put a good deal of effort into this. Instead of discussing the statistical assumptions necessary for network planning, we are presented with raw and unevaluated data. Instead of a thorough analysis of the "feasibility of a system concept", we are presented with a grandiose scheme. Buried in the pile, however, are data, which while poorly organized and presented, are necessary for prac- tical network planning. What is needed is a coherent and basic statement of the kinds of data available, of the kinds of data that are unavailable or imprecise, of the conditions under which these kinds of data hold, and of the relative usefulness of such data at varying systems levels. Perhaps it is unfair to criticize Mr. Meise for not writing this kind of book. Yet my criticism is precisely that, because he writes as though these data already exist in organized form. They don't. He has built a house of cards on air. Robert S. Taylor Book Reviews 273 Thesaurus of ERIC Descriptors, 2d Edition, Washington, D.C.: Educa- tional Resources Information Center, Bureau of Research, Office of Edu- cation, 1969. 289 pp. One of the principal problems associated with the review of a new thesaurus is that the thesaurus usually serves simultaneously to exemplify the use and misuse of the basics of thesaurus construction. The Thesaurus of ERIC Descriptors is no exception. For the purposes of this review, it is necessary to distinguish between a thesaurus and an authority list. Both are designed to improve communi- cation between the user and the information storage and retrieval system. A thesaurus is usually used in conjunction with free-vocabulary indexing (and retrieval) while the authority list must be used only with controlled- vocabulary indexing. Hence a thesaurus, in the words of the Engineers' Joint Council Guide to Indexing and Abstracting, " . .. is not meant to specify the words in which information is to be recorded, but rather to establish the semantic and generic interrelationships between such words". The indexer uses the thesaurus as a means of "enriching" his in- dexing, i.e., as a guideline for effective indexing. The searcher uses the thesaurus to aid in phrasing or clarifying his search question. In neither use is there demanded the use of a particular term in preference to any other. An authority list, on the other hand, must be composed entirely of system terminology (except for the USE-USE FOR relationships, al- though the non-preferred term cannot be used profitably as a search term) which the indexer and searcher are constrained to employ. The Thesaurus of ERIC Descriptors is, by its own admission, an au- thority list ("Only those descriptors actually used for indexing are placed in the Thesaurus .. . ", p. vii). A thesaurus may be used with either free or controlled vocabulary indexing/ retrieval; an authority list may be used only with a controlled vocabulary. It is time we started using the correct terms for these two types of communication device. Apart from the confusion as to the exact nature of the document it inu·oduces, the introduction to the Thesaurus of ERIC Descriptors does provide a good discussion of the problems of indexing and "thesaurus" de- velopment, especially concerning the need for multi-term entries. The descriptor listing, to which are added a rotated descriptor display, a descriptor group display, and descriptor scope notes, is well constructed (especially commendable is the rotated descriptor display). However, I question the value of the descriptor groups, which serve to grossly classify the ERIC descriptors, since they tend to detract from the cross-concept nature of the authority list. Finally, the formats of the various listings in this document are well done and provide a very readable and usable au- thority list. James E. Rush 274 Journal of Library Automation Vol. 2/4 December, 1969 Announced Reprints. Vol. 1, Feb. 1969. Microcard Editions. 52 pp. $30.00 per year. This journal complements Guide to Reprints. Announced Reprints lists forthcoming reprints that have been announced but not yet produced. Published quarterly, its scope includes books, journals, and other materials originating both in the United States and abroad. Each issue will cumu- late all previous issues except that following the November issue all titles that have been published will be dropped. Books are entered by author. Entries include author, title and original date of publication. Journals and sets are entered by title and include volume numbers. Each entry includes in brackets the date of the first inclusion of an item in Announced Reprints. Titles preceded by an aster- isk are those that have been published subsequent to being listed as a forthcoming title. A title that appears in the February issue, for example, as an announced title which is then published in March will appear in the May, August and November issues preceded by an asterisk. Following the November issue it is dropped. Prices are included, in some cases being in the currency of the country. Prepublication prices may be listed but the deadline is not. There is an alphabetical listing of publishers known to be active in the reprinting business, but of the 218 publishers so listed 124 did not supply Announced Reprints with titles. Among the non- respondents was Kraus Reprint Corporation, one of the larger houses. Exactly what need this journal answers is not completely clear . The Guide to Reprints provides an annual, cumulative list of books, journals and other materials that have been reprinted. As an acquisition tool it is self-evident. But since the period between the time a title is announced and the time it is actually reprinted is variable, one can only suppose that the publishers hope to fix their market by having their forthcoming titles listed in Announced Reprints. If they get expressions of interest from many libraries they may actually reprint. Since Announced Reprints gives the date of the first time a reprint title is listed, eventually librarians will learn which publishers are reliable and which are not in following through on the promise of publishing a reprint. John Demo~ Union List of Serials in the Libraries in the Miami VaUey, Sue Brown, editor. 2d edition. Dayton, Ohio: Wright State University Library, 1969. $20.00. It's hard to review a union list of serials because such a publication is obviously a very useful thing to have and to use. Being intimately con- nected with the production of a similar list for the Cincinnati area, I can Book Reviews 275 only commend the librarians of the Dayton-Miami Valley Consortium for producing this second edition in as short a time as they did. (The first edition containing 8880 titles held by 35 libraries was published in the Spring of 1968.) This edition contains the holdings of tlrree more libraries than did the first, and nearly 900 more titles are included. There are a few minor points about which one might quibble, such as the listing of the computer output on the lined side of the paper, making the pages of the published list a bit lined and grey looking; the use of corporate entries for the titles, which is O.K. if the list is used only by librarians and others used to that form of entry, but confusing to the average patron who is, I am convinced, used to looking up holdings in- formation by the running title that he picked up in a citation somewhere; listing holdings under the latest title of a periodical with notes as to the title variations over the years (although I can't complain too loudly about this, as it is the same way we are doing the Cincinnati area list, although with less information as to title changes than in this list); the use of library name "codes., that are the same as, or similar to, those used in the "Union List of Serials," which causes a great string of ODA-to run down each page. There are, naturally, a few missed cross references to the latest title as well as a few keypunching errors. These detract little from the usefulness of the volume, which should be great, especially in the area near western Ohio. The list is available for $20 from the Acquisitions Department, Library, Wright State University, Colonel Glenn Highway, Dayton, Ohio 45431. Thomas H. Rees, ]r. Current Contents; Education. 1 (June 17, 1969) Philadelphia, Institute for Scientific Information. Subscription price varies. The rise in need for librarians to build their own offprint files has in- tensified searching for current, relevant references. Current Contents; Education facilitates that search, for it reproduces contents pages of some 350 journals in the field of education and related fields. This new publi- cation includes over a dozen library journals, including the Journal of Li- brary Automation. The various sections of Current Contents have established a well de- served reputation for timeliness. Indeed, some librarians have complained that their users receive reprinted contents pages in Current Contents be- fore libraries receive the journals. Since each issue contains an author index and address directory, it is easy to request an offprint, and thereby expend minimal effort in keeping up as well as building a personal off- print collection. The subscription price can vary from $100 for a single non-educational subscription to less than $1.50 for multi-year subscriptions in groups of 200 or more. F d ·kG K"l re enc . J gour 276 Journal of Library Automation Vol. 2/ 4 December, 1969 Standardization for Documentation, Bernard Houghton, ed. Hamden, Conn.: Archon Books, 1969. 93 pp. $4.00. The editor has brought together in this tight little book an illuminating collection of six useful papers prepared initially for a conference held in Liverpool, England, in November, 1968. The announced goal of this conference was to "isolate and consider some of the areas in which the adoption of universal standards is of immediate relevance." Inasmuch as the authors are all British, the volume will have greater interest abroad than in the U.S. Nonetheless, there is universal recognition that standards in various areas of documentation are desperately needed and that a great deal remains to be done. An especially clear exposition of the British Standards Institution's work in this field is the work of C. W. Paul-Jones. He relates the methodology and the work of BSI to that of the International Standards Organization (ISO ) and touches briefly on each standards committee and its program of work. His concise outline of standards in being and in progress, and the place of each standards-involved organization in the framework of universal standards is thoroughly competent. K. I. Porter, the editor of the British Union Catalogue, touches on a variety of problems encountered in his work and discusses the potential of standards in the area of serial publications. A wryly humorous essay on standards for book production is the work of Peter Wright. He seems not very hopeful of changing the methods of book-trade production through standards, but believes in the usefulness of the effort to establish them. The essay of K.G.B. Bakewell takes up classification, cataloging, and other devices for organizing library material and providing access to it. He deplores the inchoate British de- velopment in these areas and cites considerably greater standardization elsewhere. His review of known systems is helpful. D. Martin's paper, "Standards for a Mechanized Information System," reviews the practical problems of one who has to subject information to the unthinking mind of machines. He too enumerates needed standards for coding, indexing, data elements, etc., and concludes (properly) that "It is too early to start talking in terms of solutions: standards activity is only now beginning to gather momentum." The final paper, that of John MacLachlan, is an ordi- nary how-we-do-it job, describing an abstracting service in one specific field and the local standards applied. These six papers taken as a whole constitute an informative and cogent source of information on the present status of standards work in the in- formation field. Despite the British emphasis, the case for multi-national and international standards is clearly set forth. This book should be re- quired reading for students and workers in the fields of information science and documentation. ] errold Orne Book Reviews 277 Cataloging U.S.A., by Paul S. Dunkin. Chicago: American Library Asso- ciation, 1969. 159 pp. This book is, quite simply, a survey of the development of cataloguing in America, and of the present situation of cataloguing in America. It deals with all aspects of author cataloguing, descriptive cataloguing and subject cataloguing (both subject headings and classification). The method used by the author is didactic and expository rather than critical Mr. Dunkin seeks to analyse and to display the situation rather than to arrive at startling insights or to propose radical modification. The book is addressed to "the beginning student ... the experienced cataloger ... the public service librarian ... the library administrator". It is, Mr. Dunkin says, not a "how to do it book" but a "why do it book". It is certainly true that any member of Mr. Dunkin's readership will be enlightened by be- ing shown the roots of modem cataloguing, and by having the perennial problems of cataloguing discussed in an admirably clear manner. Mr. Dun- kin does not fail to illuminate each problem, and such illumination is, of course, half way to a solution. Where he does fail, I feel, is in not pro- viding any firm answers to these problems. Perhaps in cataloguing there are no firm answers. This book seems to me, as an English cataloguer, to epitomise the "other directed" nature of American cataloguing. In read- ing this, as other American textbooks, I find a somewhat reverent attitude towards the great figures (principally Cutter), the great institutions ( prin- cipally the Library of Congress) and the "sacred texts" (the various codes). The English tradition seems to me much more "inner directed", much more concerned with what is best for the individual catalogue, much less concerned with the necessity for standardisation and consist- ency between catalogues. This is not to say that either approach has a monopoly of virtue, or that one can fault Mr. Dunkin's book on this account. Mr. Dunkin has chosen his readership and his method, and within his self-imposed limits has produced a practical and useful book. Further- more the book is written with a clarity and ease unusual in cataloguing literature. Michael Gorman Systems Analysis in Libraries, by F. Robinson, C. R. Clough, D. S. Hamil- ton, and R. Winter. Newcastle upon Tyne: Oriel Press Lin1ited, 1969. 55 pp. 15s. ( Symplegades, Number I, A Series of Papers on Computers, Libraries and Information Processing). If Symplegades once diligently guarded the entrance to the Bosphorus, it has now gratefully allowed this simple book to survive its peril. The authors explain that the title is somewhat misleading-the book 278 Journal of Library Automation Vol. 2/4 December, 1969 has nothing to do with library systems and little in terms of system analy- sis that does not relate to computerization. The two purposes of this work are the need for stressing clarity in defining objectives and for empha- sizing the extent and depth of the work involved in systems analysis. A book that could achieve such simple but difficult objectives and do it in- telligently would indeed be welcome in our discipline. This volrune, how- ever, does not obtain its objectives. It does provide something just as im- portant in that it is readable (with dashes of humor) with a simple pre- sentation of the basic tenets of systems analysis as it applies to libraries. It assumes that the reader knows nothing about systems analysis and its application to the computer. For the professional neophyte or the old graduate who has finally faced up to the realities of the future, this book should be a definite beginning point. The structure of the book and the presentation of the text contains the same simplicity as the message and book conveys. The presentation is in the form of the message. The book does contain one point of view which seems invalid. It sug- gests that systems analysis is only undertaken in connection with comput- erization. There are other shortcomings like the unexplained and unlabeled figures and the use of acronymns without explanation or definition. One also wonders about a technical book without the use of sourcing. I rene Braden On Research Libraries; Statement and Recommendations of the Commit- tee on Research Libraries of the American Council of Learned Societies; Submitted to National Advisory Commission on Libraries, November, 1967. Cambridge, Mass.: The M.I.T. Press, 1969. 104 pp. This report presents the problems of research libraries and puts forth eleven major recommendations to solve these problems. In summary the recommendations are for a national library structure presided over by a National Commission on Libraries to cope with various problems, includ- ing automation; financial support from federal, state and private sources; and study and revision of the copyright law. None of the recommenda- tions is novel. Edwin E. Williams of the Harvard University Library contributed a skillful summary of problems related to "Bibliographic Control and Physi- cal Dissemination." M. V. Mathews and W. S. Brown of the Bell Telephone Laboratories prepared a section entitled "Research Libraries and the New Technology," which discusses computers and microcopying. The discussion of library computer applications is less than helpful. The authors propose a catalog for a university library on 80 reels of magnetic tape and propose "complete resorting of the catalog." No one with any experience what- Book Reviews 279 soever in library computerization would dream, even in his worst night- mare, of such a monstrous arrangement. Yale's RalphS. Brown, Jr., has furnished an appendix, "Copyright Prob- lems of Research Libraries," that is most perceptive and informative. Brown concludes that although copyright revision must move on," the costs of using copyright works [must be] bargained out" and that Con- gress "must for a while attempt the difficult feat of standing still on a tightrope." The verso of the title page of On Research Libraries sharpens this point, for it carries the prohibition that "No part of this book may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage or retrieval system, without permission in writing from the pub- lisher." Bargaining, if it can be called that, is surely here. Frederick G. Kilgour