L&apetSLtiM, "it i e ''I s : * * ■yi\n ,4,.*. u vj/;i, iwiivvwt.t ty» ail Compatibili /j,moB| Ittformalkn. Systems s, Iltefatoe Ewbw S ! itoj l* :>l £ F yw 4^0 •■ ],!$< 'iD'efm.liirdatt oil C&i&m&'&e ■U\k S\ ■:-■■, ^ FuMicafWi 27(3 Digitized by the Internet Archive in 2013 http://archive.org/details/cooperationconveOOhend c ^ Nl ^ t ■■<■■ S*COi-C-\^ United States Department of Commerce • John T. Connor, Secretary National Bureau of Standards • A. V. Astin, Director Cooperation, Convertibility, and Compatibility Amon; Information Systems: A Literature^Review Madeline M. Henderson, John S. Moats, and Mary Elizabeth Stevens Institute for Applied Technology National Bureau of Standards Washington, D.C. and Simon M. Newman Office of Documentation National Academy of Sciences National Research Council National Bureau of Standards Miscellaneous Publication 276 Issued June 15, 1966 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C, 20402 - Price $2 Library of Congress Catalog Card Number: 66-60028 Contents Page Introduction 1 1.1. Background 1 1.2. Scope of this report 2 1.3. Definition of documentation 4 1.4. Definitions of cooperation, compatibility, and convertibility 6 1.5. Increased growth and complexity of in- formation resources 6 Early developments in coordination and co- operation 7 2.1. Historical background 7 2.2. Role of the Federal Government 9 2.3. Role of national organizations 16 2.4. Role of international groups 20 2.5. Role of conferences and meetings on documentation 22 Current operations and proposed activities 25 3.1. Background 25 3.2. Current activities in Federal agencies 25 3.3. Traditional cooperative activities 26 3.4. Dissemination and publication of secondary sources 27 3.4.1. Listings of monographic publications. 27 3.4.2. Listings of periodicals and periodical articles 29 3.4.3. Listings of reports of Government agencies 29 3.4.4. Abstracting and indexing services and publications 32 3.4.5. Clearinghouses 39 3.4.6. Copyright and translation problems.. 42 3.5. Acquisition and exchange of publications 45 3.5.1. Interlibrary loans 46 3.5.2. The Farmington and other exchange plans 47 3.5.3. Photo duplication services 49 3.5.4. Depository libraries 50 3.5.5. Cooperative exchange of data 50 3.6. Analysis and identification 51 3.6.1. Physical identification — descriptive cataloging 51 3.6.2. Content identification — subject cata- loging 54 Pag* 3. Current operations and proposed activities — Con. 3.6. Analysis and identification — Continued 3.6.3. Cooperative and centralized cata- loging 60 3.6.4. Cataloging rules and guidelines 61 3.6.5. Use of machines for description and identification 63 3.7. Svstematization and terminology control.. 66 3.7.1. Thesauri 67 3.7.2. Classification schemes 72 3.7.3. Interchange of indexing languages 77 3.8. Storage, search, and selective dissemination. 84 3.9. Standardization 86 3.9.1. Programs of Federal Government agencies 87 3.9.2. Programs of the American Standards Association (ASA) 88 3.9.3. Programs of the International Standards Organization 91 3.9.4. Programs of other organizations 93 4. Problem areas 95 4.1. General problem areas 95 4.1.1. Difficulties of problem definition 96 4.1.2. Problems of language 98 4.1.3. Consistency, reliability, and relevancy in indexing 102 4.1.4. Problems of user acceptance and human factors 104 4.2. Special problems raised by changes in the nature of the materials to be handled 107 4.3. Special problems raised by the prospects for mechanization .. 110 4.3.1. Machine language and bibliographic information 111 4.3.2. Storage requirements 113 4.3.3. Problems of input 114 4.3.4. Character sets for machine proces- sing . 116 4.3.5. Internal processing and programming languages 117 5. Implications for further progress 118 Appendix: List of references cited and selective bibliography 124 111 Cooperation, Convertibility, and Compatibility Among Information Systems: A Literature Review Madeline M. Henderson, John S. Moats, Mary Elizabeth Stevens, and Simon M. Newman The purpose of the study of the literature on which this report is based was to examine those problems in the field of documentation and in the operation of information systems which could possibly be solved or alleviated by some greater measure of cooperation, convertibility, or compatibility among systems, particularly those systems for handling scientific and technical information supported in whole or in part by the United States Government. An account is given of early developments and general background informa- tion about organizations active in cooperative documentation efforts. Current cooperative activities are then discussed in terms of dissemination and publication of secondary sources, acquisition and exchange of publications, analysis and identification, systematization and terminology control, storage and search, and standardization. General problem areas, special problems created by changes in the nature of the documentary materials to be handled and special problems raised by the prospects for mechanization are then discussed. A final section raises questions with regard to the implications for future progress. Key words : Documentation, scientific and technical information, information centers, libraries, mechanized information systems, convertibility, abstracting, indexing, cataloging, technical reports, translations, cooperative acquisi- tions, information exchange, standardization. 1. Introduction 1.1. Background There are two principal developments that have converged, since the close of World War II, to in- fluence the promotion of and participation in greater cooperation, collaboration, and coordina- tion in the handling of recorded scientific and technical information. These two developments have made the problems of such cooperation an area of sharply increased need and of challenging new opportunities to the U.S. Government, the scientific community and its professional societies, and business and industry. The first development is the large-scale involvement of the Government itself both in sponsoring and in supporting re- search and development efforts, with accompany- ing recognition of its responsibility to make public, as promptly and widely as possible, the re- sults of such efforts. The second development is the technological ad- vance that has occurred in information processing, with enormous potentialities for radically new systems and techniques. New opportunities arise from the successful use of machines to assemble, process, list, store, retrieve, and display records of various sorts ; from the development of techniques for handling scientific and technical literature by clerical and machine manipulation, and from the development of new, unconventional, and less for- mal means for publication and dissemination. A major reason for the increasing concern about possibilities for improved cooperation, converti- bility, and compatibility among information sys- tems is the problem of the so-called information explosion. Despite controversy as to the nature and extent of the steadily growing volume of scien- tific and technical information, it is indisputable that today the amount of literature that must be checked for possible pertinency far exceeds the the capacity of individuals to maintain awareness or to make full use of potentially related work. The number of scientists and engineers falls short of meeting demands; this means that there is a heavy premium on reducing duplication of efforts and conserving the technical man's time to the maximum extent possible. More effective utilization of information resources in science and technology can contribute to more effective use of scientific manpower. Such improved use of in- formation is indeed a major concern of the Com- mittee on Scientific and Technical Information (COS ATI) of the Federal Council for Science and _ Technology (FCST), The Office of the President. The present report is the outgrowth of sugges- tions advanced by the Ad Hoc Interagency Study Group on Language Compatibility in Mechanized Storage and Eetrieval Systems. At the instiga- tion of the National Science Foundation, this Study Group was established in 1961 to provide, on an informal working-level basis, a mechanism for investigation and interchange of information among representatives of agencies already in proc- ess of mechanizing their information handling op- erations, contemplating such action in the near future, or heavily involved in the interchange of information with agencies that have such mech- anized systems. During the course of early delib- erations of the Study Group it was decided that it would be well to undertake, for the guidance of the Group's members and as a source of clues to prob- lem areas and possible solutions, a review of the present state of cooperation, convertibility and compatibility among information centers. Subsequently, following FCST's establishment of a Committee on Scientific Information (precur- sor of COS ATI) its members reviewed the ob- jectives and activities of the Study Group. In a letter dated April 19, 1963, addressed to the mem- bers of the Study Group, Admiral Martell, Chair- man of the Committee, stated in part : "I am im- pressed not only with the importance of the need for cooperation, convertibilty and, if possible, ulti- mate compatibility among the many federal infor- mation systems, but also with the need for speed in isolating the many detailed problems which will need resolution as we attempt these necessary steps. I have asked Dr. Alexander to continue his chairmanship and have included your com- mittee as a technical subcommittee of the Commit- tee on Information." 1 In addition to thus co-opting the original Study Group, the Chairman approved the task of assembling a bibliography and preparing a report on the state of cooperation, convertibility, and compatibility among information retrieval sys- tems in the Government. However, the literature is not easily divided between Government and non- Government activities. "The communication problems of Government are inextricably inter- twined with those outside the Government. Both the Government and the non-Government commu- nities are concerned with the same total body of information; the progress made in each contrib- utes vitally to the other." 2 Since the lessons to be learned from a study of past actions in either case should be equally ap- plicable to present problems, a mutually accepta- ble arrangement was made to cover information systems in general. Under this arrangement, the National Science Foundation supported the intra- Government review by the National Bureau of Standards and the Office of Documentation, Na- tional Academy of Sciences-National Research Council, made provision for coverage of the extra- Government activities. A review of the literature has therefore been undertaken to determine what precedents for im- proved cooperation, convertibility, and compati- 1 Alexander and Newman, 1963 ([10], Appendix H. Note: bibliographic citations are listed in the order of the number shown in brackets in the appendix, p. 124 ff). 2 President's Scientific Advisory Committee, 1962 [465], p. 33. bility have occurred, to identify typical problems and difficulties, and to provide a realistic perspec- tive for attacks on current problems. 1.2. Scope of This Report The motivation for organization of and the name adopted by the original Ad Hoc Study Group emphasize its primary interest in machine language compatibility and, by extension, in the possibilities for cooperation, convertibility, and compatibility between mechanized systems, partic- ularly those in operation or planned by the Gov- ernment agencies engaged in large-scale operations of information dissemination. There are three major reasons, however, why these primary in- terests have needed to be broadened. First is the recognition that the problems of mechanization are but one aspect of the whole complex cycle of the generation, dissemination, and use of scientific and technical information. This recognition implies that language compati- bility at many levels (initial recording, language of announcement of availability, cataloging, classi- fying or indexing, encoding for storage, transmis- sion via communication links, and the like) affects and is affected by many different operations in the broad area of documentation. Also implied is the high probability that the problems of handling the report literature, with which the majority of the Government's information- handling agencies are most directly concerned, cannot safely be divorced from the problems of handling the traditional literature : books, serials, and monographs. In particular, the question of Government re- sponsibility for the preparation, announcement, and dissemination of information on the results of Government-sponsored research cannot be divorced from the missions of the three great Government-supported National Libraries: The Library of Congress (LC), The National Agri- cultural Library (NAL), and The National Li- brary of Medicine (NLM). Obviously, in these three libraries, the techniques for handling the traditional and the newer forms must be blended. Similarly, the network aspects of the overall prob- lem should not be neglected, which involves many small special-purpose information centers, offices, and libraries as well as the large organizations and national centers and services. The second major reason has to do with the characteristics of machines used to date and the fact that currently there are very few instances of fully mechanized systems of any considerable size or with more than quite limited operational experience. In terms of machine characteristics, there is a wide variety of machines available for prospective as well as present applications, but beyond the punched card stage, machines are seldom compatible even within the "family" of a single manufacturer. There is a true Babel of programming and machine languages; there is little or no compatibility between words or record lengths and formats ; and auxiliary equipment to achieve conversions between one type of machine- usable record and another is limited as to availa- bility, is often costly and is cumbersome and slow in performance. With respect to the limited number of mecha- nized systems from which practical lessons might be learned, it should be noted that despite the emphasis placed upon the potentialities of machine techniques over the past decade, very little real advantage has as yet been achieved. With rare exceptions, search and retrieval systems that have been mechanized in whole or in part are simply ". . . scaled up and computerized versions of punched card searching systems based on index term matching. The overall performance of such systems, in terms of the generally used precision and recall measures, appears to be very low, and not adequate to provide a thorough searching capability." 3 Large-scale integrated machine systems such as MEDLARS (Ifedical Ziterature Analysis and Retrieval -System) or STAR (Scientific and Tech- nical Aerospace Reports) are still in various stages of development and have yet to prove their comparative efficiency in overall operation, despite the promise and challenge they offer. Thus, the opinions expressed by Jahoda in 1958, Adkinson in 1960, Herbert in 1962, and an Arthur D. Little team in 1963 have almost equal pertinence today : "Mr. Gerald Jahoda reported on the results of a questionnaire survey of some 39 correlative or coordinate indexing systems now in operation, mostly located in the United States. These sys- tems are relatively small. Over 80 percent of the installations participating in the survey reported collections of 20,000 or fewer documents. The sys- tems are not extensively used, since 60 percent reported 200 or less inquiries per year. Finally, there was a strong indication from the question- naire responses that in the large majority of the installations more traditional methods such as alphabetical subject indexing or conventional clas- sifications systems might have been used to equal advantage." 4 "Thus far a large majority of the operating systems that make use of some mechanized proce- dures are located within individual industrial organizations. The subject matter encompassed by any one system is fairly homogeneous and for the most part it is chemical or biochemical in na- ture, fields in which the information itself has a more apparent structure and is more easily coded for mechanized systems than is information in other fields .... Very few of the operating sys- tems cover more than 25,000 documents, and only two or three cover more than 100,000 documents — all very small systems compared with the problem • Arthur D. Little, Inc.. 1963 [344], p. 2. 1 Stevens and de Groller 1959 [544], p. 809. of handling all scientific literature, or even all the literature of one discipline. Experience with such systems may be needed to handle very large volumes of material." 5 "The fact is that relatively few information re- trieval systems have been put to work. There are very few data available to show how various IR systems significantly excel each other, or even do better than a resourceful librarian." 6 "It may be possible to scale up a punched card searching system originally designed to work on a digital computer with perhaps 30,000 documents without too much difficulty. But scaling it up an order of magnitude further, to 300,000 documents, will in all probability so magnify and distort the existing difficulties that the feasibility of the pro- cedure becomes uncertain." 7 The third of the major reasons for extending the scope of this report beyond the "present state of cooperation, convertibility, and compatibility among information retrieval systems in the Gov- ernment" 8 was the paucity of examples found in the literature search that would directly pertain to the problems of machine language compatibility. The small number of examples of cooperation, con- vertibility, and compatibility, in the special terms of machine language and machine application compatibility, has led to a preponderance of cover- age in other areas of cooperation, especially those where future needs for greater compatibility and convertibility can be foreseen in terms of machine potentialities. For this reason, the coverage of more traditional library techniques and require- ments may seem at first glance to be out of balance; yet it well may be that improvement in the han- dling of scientific and technical information can be achieved as much by extended efforts in these areas as in specific agreements between mechanized systems. It is important to remember how in- terdependent the information generating, process- ing, and utilization processes are. The organization of material for this report has also been difficult in other respects. It has seemed desirable first to give a relatively general back- ground with respect to agencies, national and in- ternational organizations, formal and informal groups and individuals active in efforts which may improve cooperation, convertibility, or compati- bility, leading toward the ultimate goal of stand- ardization. Next, we have attempted to discuss these and similar efforts in the framework of the typical documentation and information handling operations. We have included consideration of the present and potential effects of mechanization and of typical problems and difficulties in each of these sections. However, we have also attempted to reemphasize problem areas and difficulties in a separate section, at the risk of some repetition and redundancy. 6 U.S. Senate, I960 [592], p. 104-105, report by B. W. Adkinson. 8 Herbert. 1962 [253], p. 22. 7 Arthur D. Little. Inc., 1963 [344]. p. 16-17. 8 Alexander and Newman, 1963 [9], Appendix H (italic sup- plied). In the concluding section on implications for future progress, certain somewhat controversial points have undoubtedly been raised. These re- flect points which have been made in the literature and which, in the opinion of the reviewers merit some consideration in terms of continuing concern for cooperation, convertibility, and compatibility. The authors, however, are responsible for selecting the quotations which have been included. The material covered in this report represents pri- marily information from a selective survey of the literature published through December 1963. In a few instances, however, pertinent material pub- lished later has also been included. Interesting anomalies of viewpoint, procedure and knowledge of other (but often collateral) practices run through the literature we have scanned, selected, and assembled. It is always difficult, even when teaming together authors of several different backgrounds but all with some specialized experience in the field, to avoid inad- vertent omissions and misplaced emphases in coverage and treatment. The present report was circulated in draft form to the members of the Ad Hoc Committee in Feb- ruary 1965, with a request for their criticisms and comments to be submitted by March 31. Wher- ever possible, those suggestions received for cor- rection and improvement have been incorporated. Other inaccuracies may nevertheless remain. It should be noted also that the inclusion of informa- tion on various specific activities and the omission of others is not intended to reflect an endorsement as such of those that are included nor in any sense an adverse evaluation of those that are not mentioned. The purpose of the study on which this report is based was to examine those problems in the field of documentation and in the operation of informa- tion systems which could possibly be solved or alleviated by some greater measure of cooperation, convertibility, or compatibility between systems. It might be valuable to start by defining documen- tation and by delineating what we mean by co- operation, convertibility, and compatibility among information systems. We can then enumerate the various activities pertinent to the field, point- ing out which documentation activities are sus- ceptible to cooperation and therefore are pertinent to the scope of this report. 1.3. Definition of Documentation The term "documentation" is considerably older than its present-day implications. It is reported that the delegates to the Congress of Archivists and Librarians held in Brussels in 1910 "admitted confusion regarding the use of the word docu- mention." 9 The confusion persists despite the mushrooming of the field and many new addi- tional facets. As a result little agreement on standard meanings for terms has yet been reached. Individuals and professional societies have pub- lished glossaries of terms in documentation and data processing, but little if any agreement on the use of such terms has found its way into the liter- ature of the field. However, we can quote several definitions to illustrate what is generally meant by the term "documentation." Ranganathan has traced the genesis of the term as follows : "From the turn of the present century, bibliographers began to pay attention to micro- thought (i.e., articles in current periodicals) and not merely to macro-thought (i.e., whole books). This was necessitated by large-scale promotion of research work 'in-series' and the need for com- municating nascent thought to all workers in the field as and when each little quantum of it got created. . . . Micro -bibliography and economical service of articles by photographic methods started to attract active attention about a generation ago. At that time a generic name was needed to cover these two activities. . . . This genesis of the term 'Documentation' makes it clear that the materials coming within its purview are printed and pub- lished ones arising in the course of the creation of new thought by persons working in the forefront of knowledge. . . ." 10 The letterhead of the Federation Internationale de Documentation (FID) carries the message "Documentation egal collection et conservation, classification et selection, dissemination et utilisa- tion de toute information." The term classifica- tion, while ambiguous, most often refers to the step of indexing or tagging a document with some representation of its subject content. Two addi- tional steps, those of systematization and record- ing, are of particular interest from the standpoint of cooperation, convertibility, and compatibility among information handling systems. Systemati- zation is the construction, organization and main- tenance of some form of schedule or authority list setting forth the tags or indexes, while recording is the construction and maintenance of transcrip- tions which relate the documents to the index tags. These are operations especially sensitive to the in- troduction of new techniques and mechanization which have given new emphases to the field of documentation since World War II. The new status of the field is reflected in the definition of documentation in Webster's New In- ternational Dictionary, Third Edition, as "the as- sembling, coding, and disseminating of recorded knowledge comprehensively treated as an integral procedure utilizing semantics, psychological and mechanical aids, and techniques of reproduction including microcopy for giving documentary in- formation maximum accessibility and usability." Another definition, by Helen L. Brownson, calls documentation the "art of facilitating the use of recorded, specialized knowledge through the pres- entation, reproduction, publication, dissemina- tion, collection, storage, subject analysis, organi- »Ludington, 1954 [851 1, p. 196. 10 Ranganathan, 1952 [470], p. 105-106. zation and retrieval." u In their textbook on information storage and retrieval, Becker and Hayes define documentation by pointing out what it is not: it is not concerned with librarianship so much as with the systems and tools of librarian- ship; it is not concerned with the operation of information systems so much as with their analysis and design. 12 Again, Allen Kent in his book on mechanical information retrieval considers documentation ac- tivities as the unit operations which comprise an information system: (1) acquisition of source ma- terials; (2) analysis of these documents to serve as a basis for orderly organization to facilitate the identification of wanted information on de- mand; (3) terminology and subject heading con- trol to facilitate search by the use of language based information retrieval systems; (4) record- ing the results of the analysis on a searchable 11 As contributed to the American Documentation Institute, see Wagner, 1960 [618], p. 108. "Becker and Hayes, 1903 [48], p. 43-44. medium; (5) storage of source documents, ex- tracts of documents, abstracts and bibliographic references; (6) analysis of questions put to the in- formation retrieval system and development of search strategy ; (7) conducting a search ; and (8) delivery of the results of a search. 13 If the above definitions of documentation are of service in delineating the field, then the activities of particular interest to this report are as follows : (1) dissemination, including publication of sec- ondary sources such as abstract journals, an- nouncement bulletins and indexes; (2) acquisition and exchange of publications; (3) analysis, lead- ing to descriptive cataloging and subject catalog- ing; (4) systematization, including terminology control and standardization; (5) storage, at least in some aspects such as standardization of formats, maintenance of cooperative deposits, ability to convert from one form of storage medium to an- other, ease of adding or deleting items, system modifications based on storage capacities and the " Kent, 1962 [312], 76-78; see also [309], p. 110. Table 1. Documentation Processes Process 1. COMMUNICATE A. Direct: Basic data Primary publication B. Indirect: Tables, Compendia Secondary publication Information about sources Subprocesses : 1. Identify-Select 2. Transcribe-Encode 3. Issue-Publish Process 3. ACQUIRE A. Basic data B. Primary references C. Secondary references D. Information about sources Process 2. ROUTE A. Basic data B. Primary reference C. Secondary reference Subprocesses : 1. Identify 2. Select addressees 3. Disseminate Process 4. STORE A. Basic data B. Primary references C. Secondary references D. Information about sources E. Information about needs F. Information about searches Subprocesses : 1 . Identify need 2. Locate sources 3. Order 4. Receive Process 5. SEARCH A. Basic data B. Primary references C. Secondary sources D. Information about needs Subprocesses: 1 . Organize for search 2. Organize for storage 3. Prepare and store 4. Maintain -re vise-res tore Process 6. RETRIEVE A. Selected items or values of basic data B. Selected primary references C. Selected secondary references D. Selected information about needs Subprocesses : 1. Specify selection criteria 2. Define scope, relationships of selec- tion criteria, strategy 3. Conduct search 4. Reformulate specification, iterate as necessary 5. Reject-select Subprocesses: 1 . Locate 2. Remove-reproduce 3. Prepare for use 4. Distribute 208-371— Of like; and (6) searching for documents, in the sense of "where to find what." Since the product of searching can be facts, references, or documents themselves, the interest of this report lies with what is obtained and the degree of its compatibility with the products of other systems, rather than the detailed techniques of how the product is ob- tained. We might portray documentation activities as shown in table 1. Those activities of special in- terest here, as shown by italics, are susceptible to some degree of cooperation, convertibility, or com- patibility. We can now define these terms for the purposes of this report. 1.4. Definitions of Cooperation, Compatibility, and Convertibility Cooperation involves collaborative effort or sharing between organizations in actual processing of materials for information systems or in ex- change of the products of such processing. Co- operation ranges from sharing of the load of coverage of a particular subject field by means of interorganization loans, through sharing the ab- stracting coverages and workloads, to joint effort in the establishment of word lists, thesauri, and indexing standards. An especially important area of cooperation is in the coordination of in- formation about sources — what is available and where. Systems are considered to be compatible when the results of processing in one system are immedi- ately and directly usable by other organizations having similar but not necessarily identical sys- tems. Where the products of systems in collabo- rating organizations can be used interchangeably by those organizations without special efforts or conventions, the systems are compatible. An ob- vious and early example of such compatibility is the practice of adding directly to an existing card catalog the cards supplied by the Library of Con- gress. However, it should be noted that "the most challenging problems for company, as well as for society, government, or commercial services, are those of compatibility between systems for index- ing, retrieving, and disseminating information and between the languages of those systems. There is yet, however, no meeting of the minds even as to the meaning and purpose of such compatibility." 14 Where results and products of processing in one system are usable in another system, but not im- mediately or directly, the systems may or may not be convertible one to another. Various opera- tions of transcription, transliteration, re-record- ing, re-encoding or rearrangement have to be achieved, particularly by clerical or mechanical means, in the maintenance of convertible systems. An example of convertibility is the mechanized rearrangement of a classified index listing into an alphabetized listing. Compatiblity and convertibility may apply either mutually or in only one direction between and among information systems. 1.5. Increased Growth and Complexity of Information Resources Interest in applying some measure of coopera- tion or coordination to the activities of documenta- tion has been quickened, as noted earlier, by the growing volume of publication. In addition, the increasing specialization and yet interdisciplinary nature of research areas have combined to aug- ment this same interest. In the case of scientific and technical developments, the volume of new materials is beyond the capacity of most organi- zations to acquire, issue prompt notice of accession, classify or index, disseminate, store, and subse- quently select and retrieve documents within the time scale, economics and precision necessary for efficient service and effective use. Patent offices and the legal profession compound these problems with pressures for immediate detailed documenta- tion of all past publications which, in these areas, do not lose pertinency because of age. There have been expressions of doubt as to the size and nature of the increased flow of informa- tion, and the possible extent of application of com- puter technology to the handling of this flow. 15 A particularly provocative reaction is that of Yeho- shua Bar-Hillel, who has summarized his opinion as follows: "The geometrical rate of increase in scientific and technological publications ... is commensurate with the increase in scientific and technological manpower and is in fact, nothing more than its direct result. . . . The only areas in which the [information] retrieval situation is intrinsically worsening are those which exhibit the 'cumulative effect' especially the areas of law, patent searching, and library acquisitions. But in these areas the natural remedy lies in a change of policy and practices rather than in a revolu- tionary change of the retrieval operations." w Reactions to this forthright opinion ranged from argument about the complexity of the informa- tion flow and the crisis which it engenders for librarians and information specialists to state- ments of belief that information handling facili- ties were generally keeping pace with the increased volume. An interesting point was made by Philip Abelson, editor of Science, who wrote that in lead- ing areas of scientific discovery, information re- trieval presents no problem, and there is fast and complete interchange of information by personal visit, telephone, postal correspondence, and ex- change of mimeographed preprints of papers. For "Anderson, 1962 [29], p. 114. 15 A variety of views with respect to the information "crisis" are incorporated in the report on the hearing before the Senate Subcommittee on Reorganization and Internal Organizations, 1962 [594], p. 41-49. "Bar-Hillel, 1963 [45], p. 98. those engaged in technology, who wish to apply the results of basic research, and especially for those concerned with patents, the problem is quite different and a more serious difficulty exists. 17 But given the problem of the volume and inter- disciplinary nature of scientific and technical in- formation, whether we tag it as a crisis or an ex- plosion or not, the fact remains that interest has been aroused in the possibilities for increased co- operation, compatibility, and convertibility in and among information systems. We shall review here the early steps in these directions, before discuss- ing more fully the current operations and proposed activities in these areas. 2. Early Developments in Coordination and Cooperation 2.1. Historical Background The need for cooperation by governments, in- dustry, professional societies, and individual sci- entists in the handling of scientific and technical information is not a new problem. The history of the development of science is all too abundantly marked by instances of communication failures, barriers to communication effectiveness and conse- quent replications of effort and rediscoveries of the same inventions and ideas. Disclosure of the results of research and development efforts has al- ways been prerequisite to progress in scholarship, basic scientific understanding and technological applications of the results of new knowledge. Originally, disclosure involved principally the scholarly obligations of the individual investiga- tor. The Royal Society, founded in England in the 17th Century, included among its functions the responsibility for communication and disclosure of scientific information. It is the prototype for responsibilities of professional societies and of the scientific community, then and since. Thus, until approximately the last two decades, the handling of scientific and technical information continued to be generally subsumed under obligations as- sumed by individual and professional societies to publish and disseminate disclosures of research and development results and under responsibilities for the collection and (at least to some extent) the cataloging of published literature, as assumed by libraries, museums, universities, and other institutions. Interest and concern for the effective disclosure of scientific information has had a relatively long history in the United States. We might go back to Washington's Farewell Address, in which the injunction is laid down "to promote knowledge." The Smithsonian Institution was created in 1846 as a result of a bequest from James Smithson to the United States for the establishment of a cen- ter "for the increase and diffusion of knowledge." Joseph Henry, America's most distinguished "Abelson, 1963 [1], p. 319. scientist at the time, served as Secretary for the Smithsonian for 32 years. His principles still govern it ; ". . . to facilitate in every way the pro- motion of science. . . . and enlarging the bounds of human thought." No branch of knowledge is excluded from the Smithsonian's attention. Some 10,000 scholarly books and monographs bear the Institution's imprint. Scientists and scholars in increasing numbers use its library of more than a million titles, many of them now on deposit at the Library of Congress. That venerable institution was founded at the beginning of the 19th century, when the original library — "shipped, by the Grace of God," as a bill of lading reads — arrived from London in 1801, a year after the books had been ordered by Congress. The Act of 1800 establishing the Library provided for the "purchase of such books as may be neces- sary for the use of Congress at the said city of Washington and for fitting up an apartment for containing them. The first significant material of a scientific nature came with the purchase by Con- gress in 1815 of the 7,000 volume personal library of Thomas Jefferson." 1 Today the Library of Congress includes far more than books. It houses newspapers on microfilm, motion pictures, historical items such as election posters and photographs, the finest surviving Gutenberg Bible, and the like. In all, the library is full to the bursting point with 43,500,000 items. Of particular interest to this report is the collec- tion of scientific and technical material which was started at the Smithsonian and transferred to the Library of Congress in 1866. An additional example of attempts at coopera- tion and collaboration by the Smithsonian and the Library of Congress involved library catalogs of holdings. Charles C. Jewett, librarian at the Insti- tution in its early years, saw a need for a method by which catalogs, at that time issued in book form, might be issued cheaply and promptly and without endless duplication by separate libraries. If a technique based on the "stereotype" could be de- veloped, it would be possible to assemble a national collection of stereotypes, each representing a sep- arate book, and to combine these as desired for printing the catalogs of individual libraries. "The title of every book and of each distinct edi- tion is stereotyped upon a separate plate. The author's name also stands by itself. Each plate shows at a glance the heading to which it belongs. It is obvious that these plates may be placed to- gether in alphabetical or other order, as may be desired. They are mounted on blocks, for printing like other stereotype plates. It is proposed to preserve the plates or blocks in alphabetical order of the titles, so as to be able readily to insert additional titles, in their proper places, and then to reprint the whole catalogue. By these means, the chief cost of republication (that of composi- tion) together with the trouble of revision and » Sherrod, 1962 [528], p. 35. correction of the press, would, except for new titles, be avoided." 2 "Congress, on the recommendation of the Library Committee, made an appropriation for the prac- tical testing of the plan in its application to the Library of Congress." s The resulting sample cat- alog, printed in 1854, constitutes a handsome piece of 19th century printing. Unfortunately, the plates warped in storage and the information on them was lost. Not until some 50 years later did American libraries get the benefit of a central source of bibliographic information with stand- ardized printed catalog cards based on very large collections. 4 Jewett was also instrumental in calling and pre- siding over a convention of librarians in New York in 1853, "the first convention of the kind, not only in this country, but, so far as I know, in the world." 5 We shall mention some of the actions of this 1853 Convention in section 2.5, but because the International Exchange Service also involved, his- torically, the collaboration of the Smithsonian In- stitution and the Library of Congress, we note here a communication from Mons. A. Vattemare on the subject of international exchanges and the chro- nology and coverage of cooperative efforts in this area from 1832 to that date. Several of his remarks are of particular interest : "What I aim at is, the establishment of a regular and permanent system of exchange between gov- ernments, of not only the useless duplicates of their public libraries, but everything emanating from the genius of a nation, so as to form in the capitals of the civilized nations, public international li- braries that would become a permanent exhibition of the intellectual power of each of them, a lasting World's Fair of the genius of nations. . . . Let us have a central agency on each Continent, which shall be in connection with each other to negotiate these exchanges ; let us have a monthly publication in English, French, and German, which shall pub- lish the proceedings of the agency, and the titles of the books or objects exchanged. . . . Let me close this letter by expressing my grateful acknowledg- ment towards the States and institutions of the Union, that have so readily and so nobly given a helping hand to my efforts, and tell them that, in my conviction, the time is not distant when they will reap the advantages of that generous and per- severing support ; that the little that has been done to this time is only the earnest of what is yet to come. As for the private individuals who have seconded my labors, the number is too great to mention them here, and they have already found in their conscience and patriotism the reward of their acts." 6 a "Proceedings of the Librarians' Convention [466L p. 26. 3 Ibid, p. 29. *Clapp, 1968 [116], p. 2. 5 "Proceedings of the Librarians' Convention [466], p. 13. a Ibid, pp. 53-55. 1853," 1915 1853," 1915 Under the guidance of Joseph Henry ? the Smith- sonian Institution made early provision for the international exchange of publications. Since the scientific literature collection was transferred to the Library of Congress the program has been steadily extended, until today there exist some 17,000 or more agreements with other govern- ments, research centers, laboratories, academic and other institutions (including more than a hundred in the Soviet Union) for exchanges of holdings. 7 Libraries throughout the world have been enriched by the publications received through this exchange service from many institutions in the United States and, in turn, the libraries of the United States have benefited from the publications re- ceived from the institutions in foreign countries. 8 Publications are sent to foreign countries on ex- change or as gifts by libraries, scientific societies, educational institutions, and individuals. Ship- ments of publications are received from foreign exchange bureaus for distribution in the United States. This International Exchange Service, supported by appropriations of the U.S. Congress from 1881 onward, became the official United States Agency under the terms of the Brussels Convention of 1886 for the exchange of governmental, scientific, and literary publications. 9 A final example from Smithsonian history is of interest both with respect to U.S. Government par- ticipation in international cooperation in docu- mentation and with respect to compatibility through the adoption of uniform practices. Langley, discussing the sending of U.S. delegates to the International Conference on a Catalogue of Scientific Literature (July 1895, Royal Society, London) remarked that "as to the propriety and feasibility of the United States taking part, through the Smithsonian Institution, in the pro- posed work by providing for the continuous cata- loguing of scientific literature published in the United States ... I fully concur in the view of the delegates . . . . 10 Other 19th century developments included the establishment of the libraries that have become, respectively, the National Library of Medicine and the National Agricultural Library. The National Library of Medicine "dates back to 1836, when its predecessor, the Library of the Surgeon General's Office was established." " (An interesting his- torical account of NLM, including details of early exchanges with other libraries and the earliest predecessor of today's Index Medicus, the 1840 'Catalogue'- — "quaintly inscribed on ruled paper in roughly alphabetical order" — is given by Schullian and Rogers, 1958 [507].) At least some commentators, however, suggest that "the history of the Army Medical Library 'U.S. House, 1959 [586], p. 2, statement of L. Q. Mumford. 8 Smithsonian Institution, 1964 [538], p. 65. 'True, 1946 [572]/, p. 53-54. "Rhees, 1901 [479], p. 1772. 11 National Library of Medicine, 1963 [412], p. 1. 8 as a great research institution dates back, not to its founding in 1836, but to 1865 when John Shaw Billings was an assistant to the Surgeon Gen- eral." 12 Billings is noteworthy not only for his accomplishments in the Library itself and for the Index- Catalogue but for his early appreciation of the need for union catalogs 13 and his pleas for author accuracy and specificity in the writing of titles. 14 It is of interest, especially with respect to ma- chine language compatibility, that Billings has been credited as the co-father, with Hollerith, of the notion of recording information into punched cards, which can then be sorted mechanically, and that the size of the Hollerith (IBM-type) punched card was based, at Billings' suggestion, on the then size of the U.S. dollar bill. 15 Larkey has com- mented on the Billings-Hollerith colloboration as follows : "It is not known if Billings ever thought of applying the principle to bibliographic work, but it would seem eminently fitting that it might be so utilized." 16 Another early 19th century library, involving both Government support and scientific and tech- nical information interests, is that of the U.S. Military Academy at West Point, New York. As Captain Henry Coppee remarked in 1853, "The institution which I have the honor to represent is certainly peculiar and unique — both sui juris and sui generis — in that it is under the control of the general government and that its special character is military and scientific." 17 The U.S. Department of Agriculture Library has, since its establishment in 1862, "emphasized direct service to users ... [the organic act states] that it shall be the duty of the Commissioner of Agriculture to acquire and preserve in his Depart- ment all information which he can obtain by means of books and correspondence." 18 In addition, he is to "diffuse useful information on subjects connected with agriculture in the most general and comprehensive sense of that word." 19 The 19th century also saw the first development of an abstracting service, the assumption by pro- fessional societies of responsibilities for biblio- graphic services, the founding of library associa- tions, and the first international activities in the field of documentation. For example, "As early as 1807 in the field of geology and mineralogy . . . the first abstracting service was established in Germany." 20 In 1858, the Royal Society be- gan preparation of a "Manuscript Catalogue of the Titles of Scientific Periodicals in All Lan- guages," 21 and in 1877 the Library Association of 12 Metcalf et al, 1944 [380], p. vii. 13 Schullian and Rogers, 1958 [507], p. 16. "Rogers, 1962 [490], p. 704. 15 Larkey, 1953 [3221, private communication, Scott Adams. "Larkey, 1953 [332], p. 34. 17 "Proceedings of the Librarians' Convention . . . 1853", 1915 [4661, p. 16. "U.S. Senate, 1960 [592], p. 144, statement by F. E. Mohr- hardt. 19 U.S. House, 1964 [589], p. 5. 20 U.S. House. 1959 [586]. p. 99, statement of G. M. Conrad. "Murra, 1951 [401]., p. 29-30. Great Britain was founded. 22 In 1882, The Amer- ican Association for the Advancement of Science appointed a Committee on Indexing Chemical Literature and in 1893 a Congress of Chemists held in Chicago proposed an International Committee on Chemical Bibliography to investigate coopera- tive international indexing. 23 In the remainder of this section on early develop- ments and organizational background, we shall discuss 20th century developments up to about 1960, and in a few cases we shall include even later instances as of general background interest with respect to organizations active in cooperative doc- umentation efforts. 2.2. Role of the Federal Government Burgeoning activities in scientific research and development brought about a need for extension of the responsibility for disclosure of research re- sults from that of the scholar, the professional society, and the learned institution. There was, and is, the need for disclosure as a necessary con- dition for proprietary rights to private industrial exploitation, involving patent laws and copyright requirements; disclosure as a responsibility of a publicly supported agency for dissemination to its supporters and the public; disclosure as required by contractual provisions or in return for financial support ; and disclosure as necessary to technologi- cal advancement in terms of special interests, rang- ing from the defense and security of a national government, through governmental concerns for public welfare, to the mutual advantages of com- mercial and private interests. The importance of effective handling of dis- closures of research results has received increased recognition throughout the world since World War II. In the United States, this period has been marked by : (a) Prompt assumption of Governmental re- sponsibilities for the announcement of availability of scientific and technical information stemming directly from Government-acquired material or Government-reporting of research and develop- ment progress; (b) creation of information processing opera- tions and organizations within U.S. Government agencies to handle the information output of these agencies and their contractors ; (c) continuing concern on the part of both Congressional and Executive branches of the U.S. Government with the problems of achieving more effective coordination and support of both con- ventional and newer methods of information con- trol; and (d) continuing concern with the effects of pos- sible mechanization or automation of information handling operations upon improved utilization of recorded knowledge. a Bonn, 1959 [68], p. 1445. M Murra, 1951 [401], p. 31-33. During the years of World War II, the research and development programs of the Federal Govern- ment, especially those of the Office of Scientific Re- search and Development (OSRD) and The Man- hattan Project, included provision for the release and interchange of classified information "among a veritable army of civilian scientists, but the dissemination was cumbersome because little effort was spent toward establishing and maintaining bibliographic control." 24 As early as 1944, President Roosevelt requested Dr. Vannevar Bush, Director of OSRD, to pre- pare recommendations for postwar continuation of the "unique experiment of team-work and co- operation in coordinating scientific research and in applying existing scientific knowledge to the solution of technical problems" which OSRD rep- resented. In particular, the first of four major questions was, "What can be done, consistent with military security, and with the prior approval of the military authority, to make known to the world as soon as possible the contributions which have been made during our war effort to scientific knowledge?" 25 The resulting report, "Science-the Endless Frontier," was submitted in July 1945 and served as the impetus for the organization of new infor- mation processing activities, for consideration in the Congress and elsewhere of the desirability of a Department of Science and Technology, and for many of the functions now carried by the Office of Science and Technology, The Federal Council for Science and Technology, The National Science Foundation, the Clearinghouse for Federal Sci- entific and Technical Information, and others. Examples of specific recommendations which continue to be reflected in present activities or concerns are as follows : "The Government should take an active role in promoting the international flow of scientific information." 26 "Much of the information and experience ac- quired during the war is confined to the agencies that gathered it. Except to the extent that mili- tary security dictates otherwise, such knowledge should be spread upon the record for the benefit of the general public." 27 "It is recommended that measures which will encourage and facilitate the preparation and pub- lication of reports be adopted forthwith by all agencies, governmental and private, possessing sci- entific information released from security con- trol." 28 More generally, the recommendations for a Na- tional Research Foundation foreshadow much of the structure and operations of NSF. Specifically, with respect to scientific and technical informa- tion: "Division of Publications and Scientific Collabo- ration — this Division should be charged with en- couraging the publication of scientific knowledge and promoting international exchange of scien- tific information." 29 "The Foundation should take all necessary and proper steps : "(d) To provide for and assure the most com- prehensive collection and dissemination of scien- tific and technical knowledge and information by aids to libraries, bibliographic services, translat- ing activities, etc. . . . (i) To cooperate with the Army, Navy, and civilian military research organizations for the rapid interchange of information on basic sci- entific problems of use in national defense. It should coordinate its activities wherever possible with these organizations to prevent unnecessary duplication . . . (j) To assist industry and business, partic- ularly small enterprises ... in obtaining scientific and technical information and guidance . . . " (1) To represent the United States of America in effecting better international cooperation in sci- entific activities, to assist in the freer international exchange of scentific and technical knowledge . . to help sponsor and finance international scientific, congresses or associations." 30 The work leading to "Science — the Endless Frontier" was accomplished with the aid of four committees of consultants. The reports of two of these committees are of special interest, that of the Committee on Science and the Public Welfare, Dr. Isaiah Bowman, Chairman, and that of the Committee on Publication of Scientific Informa- tion, Dr. Irvin Stewart, Chairman. The Bowman report includes as an appendix a discussion of library aids and a summary of the Committee's views on interlibrary cooperation, ab- stracting and translating services and biblio- graphic and reference services. With respect to the first, the Committee concluded that "pending the widespread adoption of really revolutionary technical aids, it will be necessary to make com- prehensive arrangements for interlibrary coopera- tion." 31 Particular problems discussed include inadequate coverage with respect to library hold- ings and the problem of reproducing European literature not available during the war, the ques- tion of translation and republication of Russian language materials, the need for cumulative in- dexes to periodicals in various fields, and the like. While the Committee did not offer specific recom- mendations, it wished "to call attention to the ex- istence of problems which, because of their mag- nitude and the large measure of centralization necessary for solution, appear to be proper sub- jects of federal concern." 32 "Heald, 1952 [244], p. 138. .„ ^ MA „ . "Letter of President Roosevelt, November 17, 1944, Busn, 1945 [91], p. vil. »Bnsh, 1945 [91], p. 17. » Ibid, p. 22. a Ibid, p. 24. 29 Ibid, p. 29. '"Ibid, pp. 110-111. "Bush, 1945 [91], p. 112. a Ibid, p. 115. 10 The Stewart Committee report, however, noting the developments already in progress in OSRD toward the establishment of a publication board, specifically recommended "the prompt establish- ment and adequate staffing within the National Academy of Sciences of the proposed board to con- trol the release and promote publication of certain scientific information. This is essential." 33 Then, in June of 1945, an Executive Order (9568) turned over to the Secretary of Commerce "operational responsibilities with respect to the declassification and distribution of government research reports kept under wraps during the war years." 34 This Executive Order was followed by a second (9604) concerning the reporting of the scientific and technical accomplishments of Axis Nations. Out of these orders came the Office of Technical Services (OTS), recently renamed the Clearinghouse for Federal Scientific and Techni- cal Information. From the original order creating a "Publication Board," derives the so-called "PB" number still in use today. The Clearinghouse was endorsed by the Federal Council for Science and Technology and is building on existing activities of OTS. Shaw reports on some of the early objectives as follows : "The work of the board is so designed as to make maximum use of all cooperating bodies and it welcomes offers of cooperation from any group which can carry responsibilities for abstract- ing and disseminating knowledge on a nonprofit basis. . . . The concept of publication under which the board operates is the broad concept of 'an- nouncing or making known' rather than the nar- rower concept of printing and distribution of mul- tiple copies. . . . The publication itself will . . . be sent to the cooperating library covering its sub- ject-matter field. Initially the cooperating li- braries are Library of Congress, Army Medical Library, and the Department of Agriculture Library . . . these libraries will file the publication by the number assigned to them by the publica- tion board." 35 In September 1950, the 81st Congress passed Public Law 776, Section 2 of which is concerned with the operation of the OTS as follows: ". . . the purpose of this Act is to make the results of technological research and development more readily available to industry and business, and to the general public, by clarifying and defining the functions and responsibilities of the Department of Commerce as a central clearinghouse for tech- nical information which is useful to American in- dustry and business . . . "The Secretary of Commerce ... is hereby di- rected : "(a) To search for, collect, classify, coordinate, record, and catalog such information from whatever sources, foreign and domestic, that may be available. "(b) To make such information available . . . through the preparation of abstracts, digests, translations, bibliographies, indexes, and micro- film and other reproductions . . ." 36 For the next several years OTS collected and processed unclassified reports supplied by cooper- ating government agencies. Then, in 1957, after Sputnik, a greater interest in Soviet scientific and technical achievements led to the establish- ment of a clearinghouse for translations. OTS has also expanded its coverage of reports from non- Government sources, particularly those result- ing from research "indirectly supported by the Government . . . (such as) grants or percentage allowances (which enable) improvements or de- velopments of research facilities . . ." 37 From its initial establishment, OTS has pro- vided many examples of cooperation in the proc- essing of scientific and technical information — intra-Governmental, between Government and in- dustry, with professional society or commercial publishers. A specific example is the program whereby American scientists and technologists visited Europe to uncover for themselves informa- tion about wartime developments there. "Victory opened the doors and the files of German factories and laboratories to American investigators. . . . The cost is trivial, no more than the salaries and expenses of investigators — they travel at no cost to the Government. . . . We are dependent on the military for billets, food, transportation, and a hundred other things, all of which they have effi- ciently provided. . . . The Office of Technical Services opens the way. . . . We impose but one major condition on investigators: they must re- port their findings fully, in writing, in the form of technical reports. All reports are made pub- licly available for the benefit of all industry." 38 Another example: "To get important docu- ments into circulation more effectively, OTS per- suaded a number of commercial publishers to issue some of them. Once every two or three weeks, a selective 'Bulletin of OTS Reports Suitable for Commercial Publication' goes to 75 publishers, in- viting them to order copies of items considered publishable." 39 Something of the scope of the "technological reparations" program of OTS may be appreciated in terms of 1947 comments: "To date over four million pages covering all fields of sciences and technology have been put on microfilm." *° Turning back to the immediate post-war period, the Atomic Energy Commission was directed by the Atomic Energy Act of 1946 (and that of 1954) , to conduct the following program: "(1) records integrate, 83 Ibid, p. 184. 84 Green, 1963 [225], p. 1. 85 Shaw [52], p. 106-107. "U.S. Senate. 1961 [591], p. 13. « Hamrick, 1963 [237]|, p. 219. b Green, 1947 [224], p. 24. >» "Distribution", 1947 [175], p. 10. «° "Technological Reparations", 1947 [567], p. 8. 11 and reproduces as rapidly as possible the scientific and technological data developed in its research and technological programs ; (2) maintains acqui- sition and exchange programs with nuclear re- search centers throughout the world ; (3) compiles the world's most comprehensive scientific inf orma- tion in the field of nuclear science ; (4) bibliograph- ically organizes, packages, and distributes for the use of all peoples this body of knowledge of the atom and its application to peaceful purposes." 41 The Department of Defense was also, of course, through the predecessor separate Departments, active in this period. The Central Air Documents Office, with a large punched card processing instal- lation, had been established at Dayton, the Navy Research Section had been established in the Divi- sion of Science and Technology of the Library of Congress, and a variety of information activities in specialized subject fields were in operation. Then, "in 1948, the Research and Development Board chartered a Committee on Technical Infor- mation. The principal act of this committee was to establish the Armed Services Technical Infor- mation Agency (ASTIA) in 1951 by consolidating the activities of the Navy at the Library of Con- gress, known as the Navy Research Section, and a similar Air Force operation at Dayton, Ohio, known as the Central Air Documents Office (CADO)." 42 The interests of the Congress in documentation and library matters, as previously noted, reach back to the beginning of the 19th century and the founding of the Library of Congress. Again, ". . . previous to the passage of the act of Con- gress establishing the Smithsonian Institution, various propositions were from time to time made to Congress, for the appropriation of the fund bequeathed to the United States . . . No one of the many plans suggested met the approval of Con- gress, until Mr. Choate proposed, and in one of his most brilliant and effective speeches advocated, the establishment of a great central library of reference and research. His bill met with gen- eral approval and passed the Senate, but was lost among other unfinished business in the lower House." 43 In 1906, the Keep Commission distributed a questionnaire on the organization and use of gov- ernment libraries, to which the Army Medical Library responded to some of the specific inquiries as follows : "On why printed cards were not used : 'About 20 languages are represented in the index- ing work of this library to several of the less known of which translations are appended. The space required can suit each case on a written card' . . . On whether the ALA list of subject headings was followed : 'For a highly specialized collection the subject headings of the A.L.A., though excellent for a general library, would be inadequate.' " 44 The interest of the Senate Committee on Gov- ernment Operations in the development of a co- ordinated program for dissemination of scientific information began when the committee was first established in the 80th Congress. "Immediately after its creation . . . the committee . . . held hearings on a bill (S 493) ... to provide for the coordination of agencies disseminating technolog- ical and scientific information and for the more efficient administration of an information ex- change program." 45 But instead of S. 493, a bill was passed which created the National Science Foundation (NSF) . This 1950 act (P.L. 81-507), reflected recognition by the Congress "of a new and enchanced role of science and technology in public affairs — recognition by an expressed intent : to promote the progress of science ; to advance the national health, prosperity, and welfare ; to secure the national defense." 46 Paragraph 5 of the act is especially significant: "To foster the interchange of scientific information among scientists in the United States and foreign countries . . . ." 47 In 1958, the National Defense Education Act provided for the establishment of the Science In- formation Service in the NSF, as follows: "The National Science Foundation, through . . . (the Science Information Service) . . . shall (1) pro- vide, or arrange for the provision of, indexing, ab- stracting, translating, and other services leading to a more effective dissemination of scientific in- formation; and (2) undertake programs to de- velop new or improved methods, including mecha- nized systems, for making scientific information available." 48 On December 30, 1958, the Science Information Council advisory to the Office of Science Infor- mation Service (OSIS) was established. The Council included the Librarian of Congress, the Director of the National Library of Medicine, the Director of the National Agricultural Library, the head of OSIS, and 15 members from the fields of science, librarianship, and documentation, and from the lay public. NSF also established the Fed- eral Advisory Committee on Scientific Informa- tion, which held its first meeting in January 1959, and considered the problem of dissemination of unpublished research reports and the need for ex- panded indexing services for foreign technical literature. Executive Order No. 10807 of March 13, 1959 established the Federal Council for Science and Technology "to promote closer cooperation among Federal agencies in planning their research and development programs, and to recommend ways in which the Federal Government can assist in ad- 41 U.S. Senate, 1960 [592], p. 36. "U.S. House 1959 [586]. p. 43, statement of B. 43 "Proceedings of the Librarians' Convention . . [466], p. 19. G. Huff. . 1853," 1915 *» Schullian and Rogers, 1958 [567], p. 29. "U.S. Senate, 1960 [592], p. 1. 40 U.S. Senate, 1961 [591], p. 10. 47 Par. 5, The National Science Foundation Act of 1950 (P.L. 81-507). 48 Title IX, National Defense Education Act of 1958 (P.L. 85-864). 12 vancing and strengthening the Nation's scientific effort as a whole". 49 The Order called specifi- cally for the National Science Foundation to pro- vide leadership in the effective coordination of the scientific information activities of the Federal Government and to improve the availability and dissemination of scientific information. In carrying out the various tasks and responsi- bilities assigned to it, the NSF "is generally pro- ceeding on the assumption that more can be gained by close cooperation with, and in support of, exist- ing scientific information services in the United States, both public and private, where they are functioning effectively, than by direct Federal op- eration of such services. The scientific informa- tion services rendered by many of the scientific so- cities and professional institutions to the scientific community are world famous for their quality. We believe it is essential that the Federal Govern- ment continue to cooperate with, and assist, such private groups in the achievement of long-range solutions to scientific information problems." 50 In summarizing the activities of OSIS in its first several years of operation, Adkinson reports that a six-point program has been developed by the National Science Foundation for improving the dissemination of scientific information, par- ticularly among U.S. scientists and engineers : "First, improvement in policy making, program planning, and coordination in scientific informa- tion activities is being fostered both within and be- tween Government agencies. "Second, a sizable research and study program is being supported looking toward the develop- ment of new and better methods of processing, disseminating, storage, and retrieving scientific information . . . "Third, scientific societies are being encouraged and supported by the Federal Government in as- suming greater responsibility for the effective dis- semination of scientific information within their own disciplines . . . "Fourth, closer cooperation and coordination is being fostered among Government agencies hav- ing like or similar scientific information pro- grams . . . "Fifth, better coordination is being fostered in American participation in international organi- zations that are oriented to documentation ac- tivities . . . "Sixth, action is being undertaken to stimulate effective educational and training programs in scientific documentation." 51 As a result of these and other programs, it has been claimed that "O.S.I.S has probably served as the most influential single force for progress in scientific and technical information throughout the world." 52 •"•U.S. Senate, 1961 [591]/, p. 15. 60 U.S. Senate, 1960 [592], p. 25, comments by A. T. Waterman. "1962 [51. p. 50. "Cahn, 1962 [94], p. 28. Other examples of the U.S. Government activi- ties during the first 15 years succeeding the close of World War II include the "financial sponsor- ship by agencies of the publication of new research journals by appropriate learned societies," such as the Air Force Office of Scientific Research in the field of fluid mechanics and the AEC in the field of reactor technology, 53 the DOD Directive of March 4, 1952 establishing the Armed Forces Medical Library as a national library for medicine and related sciences, and the subsequent transfer- rence of the latter to the Department of Health, Education, and Welfare with the name change ap- propriate to its mission. To cope with its information handling problems, the Central Intelligence Agency has since the late 1940's developed an information processing center which "comprehensively indexes and stores that information which is collected and, as a service of common concern, renders daily support to analysts at work in all parts of the U.S. Government's intel- ligence community." 54 In 1954, in its consideration of a Department of Commerce budget request, the U.S. Senate Appro- priations Committee had directed that the Depart- ment ". . . make an aggressive, thorough investi- gation as to the possibility of mechanizing the searching operations . . . '[of the U.S. Patent Office] ." 55 This directive resulted in the conven- ing of a committee headed by Dr. Vannevar Bush which recommended in part that a research unit be established in the Patent Office and that it undertake a joint program with the National Bu- reau of Standards to develop techniques specifi- cally for patent searching operations (1954 [596] ) . Turning now to cooperative efforts among the various Government agencies, we note first that during the period 1946-1948 the Army Medical Library entered into a cooperative cataloging agreement with the Library of Congress whereby "the cards for all medical titles cataloged by the two libraries were published in a medical card series." 56 Far more extensive, however, were the pre-VE Day concerns of the defense-intelligence communities with problems of acquisition of addi- tional and declassification of previously acquired material relating to foreign scientific and tech- nological information. "The burden of federal procurement throughout the war . . . was placed upon two principal sources of supply — the foreign service and an interdepartmental committee created for the express purpose of acquiring for- eign publications." 57 Further, "during the past few years the govern- ment has itself, through the Office of Strategic Services, the Interdepartmental Committee for the Acquisition of Foreign Publications, the Depart- ment of State, and the Library of Congress, been G3 Herner and Herner, 1959 [254], p. 195. M U.S. Senate, 1960 [592], p. 63. 65 Ibid. p. 75. M MaeDonald, 1953 [364], p. 67. "Humphrey, 1946 [276], p. 99. 13 the principal promoter of the acquisition of Eu- ropean (as well as other) book materials, and has made, through the Alien Property Custodian's re- publication program, many important foreign publications available in photofacsimile to librar- ies generally .... "At the June [1945] meeting of the Association of Research Libraries, the question was raised whether libraries generally might not share the use of the channels enjoyed by the Library of Con- gress. On August 4, 1945, Mr. MacLeish wrote to Dr. Evans stating that : 'The Department of State agrees with the Library of Congress's view that the national interest is directly affected by the holdings of the many private research libraries. It would, therefore, interpose no objection in prin- ciple to the employment of federal government fa- cilities to assist in maintaining their specialized collections where normal channels of acquisition are inoperative. It is believed, however, that cer- tain basic understandings should be made clear at the outset. The department would wish to be assured that the private libraries had agreed upon and carefully planned a program of cooperative buying . . . .' " 58 Shortly after the war, then, the Department of State requested the Librarian of Congress to ex- plore new means for coordination of the foreign acquisitions activities. As a result, it was recommended to the Secretary of State that consideration be given to "the estab- lishment of a permanent Interdepartmental Com- mittee on the Acquisition of Library Materials within the framework of the interagency intelli- gence group which the President had requested him to form." 69 In late 1946, the area of cooperative acquisition was the topic of a conference of college librarians. Clapp, in his introduction, said : "As a sort of pre- cursor of a general plan, there has been developed a special project which has had as its object the procurement and distribution of recent European library materials," 60 and Peiss reported on the Library of Congress Mission, stating that "the number of volumes shipped from Germany and Austria . . . now exceeds one million." 61 Boyd, who was a co-father of what would emerge as the Farmington Plan, remarked at the same conference that "some of us have been privileged to inspect the vast mechanism of the cooperative acquisitions project. More than six thousand wooden packing cases full of books, gathered from many repositories in Europe, trans- ported to America, opened, divided into categories, correlated with the system of priorities established by the committee and dispatched to 115 participat- ing libraries, makes an impressing spectacle." 62 The Downs committee, which established these 68 Clapp, 1946 [115], p. 130-131. "Humphrey, 1946 [276], p. 103. 60 Clapp, 1947 [113], p. 100. 81 Pelss, 1947 [452], p. 116. •"Boyd, 1947 [75], p. 109. 14 priorities among participating libraries for the distribution of materials found by the Cooperative Acquisitions Project, was composed of representa- tives of ALA, the Association of Research Libraries, the American Council of Learned Societies, the Joint Committee on Importations, the National Research Council, the Social Science Research Council, and the American Council on Education. The Interdepartmental Committee itself was of interest in several ways. First, it provided an instructive example with respect to the roles of members of such interagency groups : "For mem- bers of the committee to conceive of them- selves, or for their agencies to conceive of them, solely as representatives of the interests of the governmental bodies to which they are attached, would render most difficult the primary task of attaining broad consideration of federal acquisi- tion policy. The situation demands, instead, the continuous deliberation of individuals whose re- sponsibility and chief interest lies in substantive fields of knowledge not necessarily encompassed by the rigid framework of governmental adminis- trative structures." G3 Second, the Committee provided a stimulus to other more or less formal mechanisms for inter- change of information and for cooperation among Government agencies. Taube, for example, from 1948 onward promoted collaboration between The Science and Technology Division, LC, and the information divisions of other agencies. Taube (1952 [560]) also has described arrangements made by AEC to receive photostats of all articles on selected subjects abstracted by either the U.S. Department of Agriculture Library or the Army Medical Library. Such efforts bore fruit in these and other govern- ment agencies, in the Group for the Standardiza- tion of Information Services (GSIS) which was organized in the early 1950's. The agencies represented in the Group included the Technical Information Service of the Atomic Energy Com- mission (AEC) ; the Division of Research In- formation of the National Advisory Committee for Aeronautics (NACA), now a part of the Na- tional Aeronautical and Space Agency (NASA) ; the Central Air Documents Office (CADO) of the Air Materiel Command; and the Navy Research Section of the Library of Congress (LC) the lat- ter two being eventually merged as ASTIA. Working-level personnel of the agencies repre- sented in the Group agreed upon a common format for catalog cards in the hope that they might be interchangeable among and directly usable by the agencies in their respective card catalogs. The card, which had been devised at LC by Taube, was a marginal-type 3x5 card having a vertical line about two inches from the right edge ; the citation and an abstract appeared in the body of the card, and subject headings and other filing points in the «» Humphrey, 1946 [276], p. 103. right-hand margin. The cards could also be em- ployed in the preparation of abstract bulletins, bibliographies, and cumulated indexes ; pasted into sheets, with the margins covered, they formed two- column pages of entries, each of which consisted of a citation and an abstract. This technique was in fact used in the preparation of Technical Informa- tion Pilot (TIP), of the Navy Eesearch Sec- tion. "The TIP . . . utilizes the left half of the GSIS catalog card for each entry and makes addi- tional typing for the journal unnecessary except for the typing of source and subject indexes for each issue." 64 Other GSIS activities, for example, with respect to possibilities for standard practice in the citing of corporate authors and for the development of a common set of subject headings, were significantly less successful. 65 The interesting point in these attempts at cooperative processing was that the working-level personnel got together, established common aims and at least some common goals and tried to operate in a cooperative and coordinated manner. It is also interesting to note that several non-Governmental agencies, the editorial office of Meteorological and Geoastrophysical Abstracts and a number of Department of Defense contrac- tors, adopted the right-hand margin card. Another early example of cooperative Govern- mental agency cooperation to improve the dissem- ination of information about research accomplish- ments and research progress was the short-lived periodical, Federal Science Progress. The first issue, for February, 1947, stated in a foreword that : "This is a magazine for the businessman. It is published to bring him, each month, a report on the scientific and technical activities of his Gov- ernment . . . Government research files are bulg- ing with reports. They contain millions of pages of scientific and technical information seized in enemy countries, as well as information given us by friendly foreign governments. Federal Science Progress proposes to skim the cream from this mass of material and to report to the businessman what is available, and where." 66 After a relatively small number of issues, publi- cation of Federal Science Progress was discon- tinued. As the endpiece of its last issue, there is a facsimile of a Department of Commerce press release dated March 7, 1947, stating in part: " 'Federal Science Progress' was started with the February issue as a medium to review and to call attention to the nonsecret scientific reports avail- able in the Government of work being carried on with federal funds . . . The magazine . . . was regarded after three month's experience as not being adapted to the purely service functions of the Department . . . Criticism developed from publishers with general circulation in the scientific field on the ground that the magazine represented potential competition, and overlapped the field of private publications. Although early Department surveys had indicated that direct competition would not develop, the views of publisher groups were taken into consideration in making the decision." 67 In the first issue of this ill-fated periodical, the ill-fated Chemical-Biological Coordination Center received early appreciation : "Now there is a cen- tral clearinghouse to record all new substances, as well as those already discovered, to keep a system- atic file on their most important properties, and to point the way to further research on their potentalities. This tremendous job has been undertaken by the Chemical-Biological Coordi- nating Center, recently set up by the National Research Council . . . ." 68 During the war years, OSRD had sponsored the testing of chemical compounds for biological activity, especially for effectiveness against in- sects and rodents. After the war, its Insect Con- trol Committee was transferred to the National Eesearch Council, where with financial support from the Army, the Navy and the American Can- cer Society it emerged in 1946 as the Chemical- Biological Coordination Center. The initial ob- jectives were : "to make information accessible to scientists and to facilitate the study of relation- ships between chemical structure and biological activity ... to sponsor a screening program to facilitate testing of chemicals on a variety of plants and animals ... to prepare and publish reviews, and to sponsor symposia." 69 The essen- tial first steps were to develop codes for both the chemical and the biological information. In a study made for the Air Force Office of Sci- entific Research, Dougherty has explored the "rise and fall" of this organization. The experiment was an interesting one, especially in the develop- ment of the codes. However, many of the reasons for failure are instructive in terms of the general problems of improving utilization of scientific and technical information through cooperation, con- vertibility, and compatibility. First, the questions of funding. By the end of CBCC's first five years, budget requirements were in excess of the planned financial support. "In the beginning of the Center apparently the spon- sors believed in its value; now, however, it has become more expensive than had originally been estimated, therefore, the NRC wants to reappraise its value and cost. The Research Council is inter- ested in furnishing a home for the CBCC as an interesting and perhaps very important experi- ment but only as long as there is adequate support and interest." 70 There was, second, the problem of alternate or supplemental sources for stable financing. At a M Jackson, 1052 [291], p. 96. "Taube, 1952 [560], p. 22-23. 00 Federal Science Progress 1, No. 1, 3 (1947). 67 Federal Science Progress 1, No. 5 (1947). « "Chemical Center," 1947 [110], p. 15. 60 Dougherty, 1963 [177], p. 4-5. *° Minutes, 1952 meeting of representatives of sponsoring agen- cies, quoted, Dougherty, 1963 [177]) p. 9. 15 1952 meeting, "a number of those present sug- gested the possibility of approaching industry for support . . . Others . . . pointed out that com- panies would not want to cooperate until they found something they could utilize, or, that indus- try would not be willing to support the Center until provisions had been made to insure indi- vidual companies that their patent rights would be protected." 71 A third problem was that of relative breadth and depth of coverage, especially as CBCC moved more in the direction of emphasizing service func- tions. In 1955 a special committee considered, among other problems, the scope of the activities. "The term 'scope' was used to denote two concepts. First, that the amount of literature coverage be increased, that is, the number of journals coded in the various disciplines be increased, and second, that the size of the organization, itself, be ex- panded in order to broaden the scope of services offered. There were several people who advo- cated expansion. On the other hand, there were others who recommended that instead of attempt- ing to cover so many disciplines that the center should concentrate its efforts and restrict coverage to only a few disciplines in order to achieve depth. Finally there were a few who believed that the Center needed both breadth and depth of cover- age in order to achieve its objective . . . ." 72 Finally, in 1957, NAS/NRC terminated the CBCC operations. Dougherty reports represent- ative reactions as follows: "Sentiment ran the gamut from thorough disgust with the Academy's decision to complete approval. . . . The Center received a number of letters from scientists and organizations that had received information from the Center . . . Although the letters, about thirty in number, could not be termed a representative sample, they indicated that some scientists had benefited from the Center's services and regretted its passing." 73 Further "some scientists believed that ... if activities were terminated somebody else would have to restart a similar operation," "the center was ahead of its time," "the center had demon- strated that given adequate financial support and scientifically qualified personnel large masses of published and unpublished data could be con- trolled by machine methods," "the nature of the objectives had forced the staff to spread itself so thinly in its efforts that the ultimate result was over-all dilution of programs." 7i The most sig- nificant achievement was generally felt to be the development of the NEC Chemistry and Biology Codes. "An important early contribution to informa- tion retrieval was made by the Chemical-Bio- logical Coordination Center . . . There were several outgrowths as a result of the early pioneer- 71 Dougherty. 1963 [177], p. 9. ™ Ibid, pp. 10-11. « Ibid, pp. 13-14. 74 Ibid, pp. 15-16. ing work of the CBCC . . . Among them was the work of Maloney, at Fort Detrick, who developed a medical and chemical system patterned after the CBCC work. Welt developed one of the CBCC ramifications which has subsequently become the Cardiovascular Literature Project. A somewhat related project dealing with carcinogenic effects of chemical compounds is being carried out at The Cancer Chemotherapy National Service Cen- ter .. . ." 75 This example has been reported at some length because the CBCC involved a cooperative venture in information services in terms of both Govern- mental and non-Governmental sponsorship, it en- listed the collaboration of outside scientists in various subcommittees which dealt with the de- velopment of the codes, it attempted to develop an integrated system for storage, search, and retrieval using machine techniques. Moreover, it is possi- ble that instructive lessons may be learned from the reasons for failure. Dougherty summarizes the four points con- sidered to be the major reasons as follows: "The first, as stated by the NRC-NAS, was an inability to attract adequate and stable financial support,.. The second, according to a number of scientists, was the failure to limit its objectives and scope of operation . . . The third point was that the Cen- ter had not been able to develop an integrated storage and retrieval system. While they had demonstrated an ability to code and store data, the retrieval portion of the system had been neglected. As a consequence the Center had not been able to show a capacity to retrieve stored data. Finally, one scientist stated that the Center had not been able to demonstrate the uniqueness of its services. In his opinion, the only unique features produced were the two codes and the design of a variety of code sheets." 76 2.3. Role of National Organizations In addition to early Government activities lead- ing to cooperation and coordination of scientific information activities, other steps have been taken by organizations interested in such developments. Associations of librarians; professional societies in fields related to library science and in fields re- lated to science, technology, law and communica- tions; industrial organizations, and formal and informal groupings of many of these bodies have individually and in mutual cooperation con- tributed to the development of documentation practices. For example, "the Joint Libraries Committee on Fair Use in Photocopying . . . which represents the American Library Association, the Association of Research Libraries, and the Special Libraries Association, was established in 1957 to work on copyright problems which arise when libraries 75 Marden & Koller, 1961 [370], p. 11. 70 Dougherty, 1963 [177 J, p. 17. 16 photocopy materials in their collections for their users. . . ." 7T Among the librarians particularly active in documentation matters have been the members of the Special Libraries Association (SLA) , "formed in 1909 by librarians of business, professional, governmental, and industrial organizations to promote the collection, organization and dissemi- nation of information in specialized fields and to improve the usefulness of special libraries and information services." 78 One instance of SLA's interest in cooperation in documentation was the session held in conjunction with the annual con- vention of 1959, which had the theme "Interna- tional Cooperation in Documentation." Papers were presented on documentation services in Great Britain, the Netherlands, West Germany, South Africa, Japan, Canada, and Latin America, as well as on two documentation agencies in the United States. Other instances of cooperative activities on the part of SLA are the maintenance of a collection of specialized classification schedules, contributed from other groups in the United States and abroad and the establishment, in 1953, of a national "pool" of translations at the John Crerar Library which, in 1959, entered into a cooperative program with OTS. 79 The American Documentation Institute (ADI) was formed in 1937 as a society of organizations rather than of individuals. Its original major in- terests were "in techniques for more effective use of microfilm and questions related to copyright . . . based on the goal which ADI had set for it- self ... to make scholarly material readily avail- able to researchers . . . . 80 As steps to implement the goal, a network of microfilm laboratories, a coordinated interlibrary loan system, and an auxil- iary publication program were envisioned. The latter program, which began in the U.S. Depart- ment of Agriculture Library in 1934, 81 still oper- ates through the Photoduplication Service of the Library of Congress. In 1952, ADI changed from an institute of so- cieties to an organization composed mainly of in- dividual members, although institutional member- ship is still also allowed. The interests of the so- ciety and the activities to which it has turned its attention have broadened considerably. By 1957, ADI had working committees on Bibliography, on Research and Development, on Education of Documentalists, and on Cataloging and Classifi- cation. Until 1959, ADI also served as the U.S. repre- sentative to the International Federation for Doc- umentation (FID). When that international organization was incorporated under Belgian law, in August 1959, and increased the financial obli- " Scientific Information Notes, Vol. 3, No. 4, p. 20 (1961). "Towner, 1953 [5711, title page. "Gingold, 1961 [217], p. 15; U.S. House, 1959 [586], p. 19. 8 »Adkinson, 1964 [4], p. 388. 81 Mohrhardt, 1959 [391 L p. 30. gations of national and international members, A.DI requested the National Academy of Sciences- National Research Council to undertake the re- sponsibility. The Office of Documentation, NAS/ NRC, was assigned the task of establishing the U.S. National Committee for FID, and "atiny membership organization . . . was thus relieved of the burden of this representation." 82 An example of the continuing drive of ADI to explore and to apply new techniques for coping with documentation problems was shown in the preparation for its 1963 annual meeting. Each registrant for the meeting was given two volumes of preprints of "short papers" (not exceeding two printed pages) , the text of which had been printed from type set automatically with the aid of elec- tronic information processing equipment, in an interval of three weeks or less before the meeting. (Luhn, 1963 (354) : Figure 1 is a facsimile of the frontispiece of a special edition of Volume I) . FIRST BOOK OT TECHN1CM ARTICLES TYPE-SET BY COMPUTER Th!i It No. 68. of I 00 coplet of a ipeclal edition of thit book, prepared at a memento and a> a token of recognition to thoie who were Involved in Itt creation and who are here identified by their tlgnahireti g/?&l~J* fj$i/)fr4^tL Ccuntortlgnod Chicago, 111. October 6, 1943 TikX PRESIDENT. AMEHICAN OOCTjMErtjTATION INSTITUTE The bibliographic information for the approxi- mately 600 papers was keypunched and processed on an IBM-1401 computer to produce (1) the table of contents, (2) a permuted title or key word-in- context (KWIC) index, (3) an author index to the short papers, (4) a citation index to the biblio- graphic references cited in the papers, (5) a KWIC index to the titles of these references, (6) a "bibliography" or listing by the codes assigned 82 Council on Library Resources, 1960 [142], p. 6. 17 to the citations, and (7) an author index to the ci- tations. The machine-readable form of the text is available for research purposes. In appointing the U.S. National Committee for FID, the Office of Documentation, NAS/NRC, provides representation not only for documenta- tion and scientific information specialists, but also for scientific societies. Many of these societies have their own groups and programs in the scien- tific information field. The Division of Chemical Literature of the American Chemical Society (ACS) dates at least informally from the society's meeting in the Spring of 1943. At that meeting a paper by Egl- off, Alexander and Van Arsdell, "Problems of the Scientific Literature Survey" (1949 [184]), was presented before the Division of Chemical Educa- tion. The presentation inspired the organization of symposia on technical library techniques, the first of which was held at the ACS Fall meeting in 1943. This activity in turn led to the formation of the Chemical Literature Group of the Division of Chemical Education. The Group sponsored ses- sions at subsequent national meetings, dealing with the subjects of technical library operation, chemi- cal nomenclature, translation of foreign publica- tions, and the like. In September 1948, the ACS Council approved the petition of the Group to be given the status of a division, and the Division of Chemical Literature came into existence. Its first official program was presented at the Society's Spring 1949 meeting. The Division continues to devote itself to documentation problems in the field of chemistry, sponsors technical sessions, maintains operating committees for various spe- cial areas and is responsible for technical content of the ACS publication Journal of Chemical Doc- umentation. In addition to its sponsorship of the Gordon Research Conferences, the American Association for the Advancement of Science (AAAS) has es- tablished the Section on Information and Com- munication (Section T), which developed from a series of information conferences which had been held during the Association's annual meetings be- ginning in 1951. The conferences were variously called Operation Knowledge (1951), Conference on Scientific Editorial Problems (1952-56), and Conference on Scientific Communication (1957- 1961). The Section was organized formally in 1962. It has provided a forum where interested societies and individuals may "discuss documenta- tion problems and work toward their solution." 83 The Special Interest Group on Information Re- trieval (SIGIR) of the Association for Comput- ing Machinery (ACM) was organized in 1962 to satisfy certain specialized needs of the professional community served by ACM. These include, (1) advancement of the application of machines to the storage, retrieval, and dissemination of informa- 83 Parkins, 1962 [447], p. 536. tion; (2) advancement of information retrieval theory, programming, file preparation, searching strategy, output schemes, information retrieval system evaluation, and development of equipment best suited for these tasks; and (3) the exchange of information on new developments in these areas. Two groups actively engaged in cooperative and coordinating studies of documentation problems are the National Federation of Science Abstract- ing and Indexing Services (NFSAIS), and the Classification Research Study Group (CRSG). The NFSAIS was established in 1958, when the National Science Foundation "provided funds and staff work for a conference of major U.S. scien- tific abstracting and indexing services. The ma j or accomplishment of the conference was the forma- tion of the National Federation of Science Ab- stracting and Indexing Services which strives to coordinate and improve the work of the various services and to encourage the development of ab- stracting and indexing services for those special- ized subject fields not at present covered by such services." 84 The Federation now numbers 14 major abstract- ing and indexing service organizations. Its ac- tivities include the preparation of a union list of periodicals covered by the cooperating services, analysis of the scientific periodicals produced around the world, and the like. Its working groups are the Index Editorial, Management Serv- ices, Manufacture and Production, Material Pro- curement, and Research Groups. The Classification Research Study Group is an organization of persons interested in the develop- ment of classification schemes and often actively engaged in research or studies of classification sys- tems and applications. The group usually meets, in conjunction with meetings of related societies, and conducts cooperative projects dealing with the theory and structure of classification schemes. It collaborates also in the maintenance of the SLA Special Classifications Center. In addition to these specialized groups, the fol- lowing organizations are actively interested in cer- tain aspects of cooperative documentation: the American Institute of Physics, with its Docu- mentation Research Project; the American Chem- ical Society which in addition to the Division of Chemical Literature has an active research pro- gram on abstracting and indexing and the total publication process and which "since its inception in 1876 . . . has been deeply concerned with the need to handle effectively the mass of chemical knowledge generated by American scientists," 8B and the American Institute of Biological Sciences. The Biological Sciences Communication Project, for example, studies the flow of information from, producer to user, the extent of use of the foreign literature, citation practices, and the like. One of 84 U.S. House, 1959 [591], p. 67, statement by A. T. Waterman. 85 U.S. Senate, 1962 [594], p. 24, statement of A. H. Emery. 18 its most important tasks is the "identification, ex- amination, and classification of all the world's journals in the field of biomedicine." 86 Still other organizations interested in coopera- tive efforts in documentation include the National Microfilm Association, concerned with all aspects of microfilm reprography ; the Association of Re- search Libraries, whose Automation Committee proposes to take a major role in developing a na- tional system of bibliographic automation; the American Society for Metals with its Documenta- tion Committee and its collaboration with Western Reserve; the Engineers Joint Council and its In- formation System Committee, and the American Library Association which in 1922 established a Committee on Bibliography that would partici- pate "in the growing movement to get together all going enterprises in biblography for practical co- operation," 87 and which has sponsored since 1959 a_ Library Technology Project "designed to pro- vide librarians with authoritative information and guidance of all phases of equipment, supplies, ma- chines, and systems used in library operations." 88 An example of early interindustry cooperative concern in the technical documentation area is the American Petroleum Institute : ". . . Inthel940's the_ considerable amount of duplicate work going on in processing a central core of published petro- leum information, and in developing the methods needed for this processing, caused much concern. It seemed logical that much of this work could be done more economically, and on an even larger scale, by a central operation. Each company would still have to cover a residue of documents of particular interest to itself. Interest was aroused among petroleum research managements, and a proposal to begin centralized abstracting was eventually put before the American Petro- leum Institute . . . ." "[In 1954] the API Central Abstracting Service began its operations by publishing a weekly ab- stract bulletin covering the literature reported in the 100 sources deemed basic to petroleum refin- ing. Classified sets of abstract cards were soon made available, following development of an ac- ceptable classification system." 89 _ The interests of some of the nonprofit organiza- tions and private foundations in the United States should also be noted. In his 1945 report to the President, Bush pointed out that "both the Library of Congress and the Army Medical Library oc- cupy leading positions in their fields. Yet those two Government institutions still have to look to private sources for much of their support, espe- cially for projects involving experimentation with new methods. Two foundations alone have con- tributed over half a million dollars to the Library of Congress in the past few years." 90 80 Shilling 1963 [531]/, p. 206. 87 Richardson, 1928 [480],, p. 43. 88 Scientific Information Notes 2, No. 2, 3 (1960). 69 Weil et al., 1961 [635] , p. 57. "Bush, 1945 [91], p. 113. First and foremost among such organizations, partly because of its quasi-Governmental yet in- dependent character, partly because of its sponsor- ship of work involved in the preparation of this report, and partly because of its explicit assump- tion of responsibilities for international collabora- tion in the field, is the National Academy of Sci- ences/National Research Council. NAS/NRC has been instrumental in such other activities as the planning and support of the International Con- ference on Scientific Information, the establish- ment of the Office of Critical Tables, and provision of financial support and/or "homes" for special- ized information centers such as CBCC and the Prevention of Deterioration Center which "is a nonprofit, scientific organization maintained jointly by the three United States armed services, by means of an Office of Naval Research contract, under the operating supervision of the National Academy of Science-National Research Coun- cil." 91 Another example of NAS/NRC service in U.S. representation to international groups was the U.S. National Committee for the International Geophysical Year. We might note in passing that IGY had been preceded by an International Polar Year "in which several countries interested in the compilation of scientific data of the Arctic regions, pooled information and techniques. This cooperation gave added value to the methods of the observation and presentation of the data col- lected." 92 IGY activities are also to be noted in terms of uses of advanced documentation techniques. Thus, computer programs were prepared to compile a "Tabledex" index for the IGY bibliography com- piled by the Library of Congress (Zusman et al., 1962 [660]) and the World Meteorological Or- ganization published the meteorological data ob- tained in the form of micro-cards (Giinther, 1962 [232]). In addition to its contributions over the years to the Library of Congress, Army Medical Li- brary, and others, the Rockefeller Foundation's interests are exemplified by its initial financial sup- port for the American Book Center, (now the U.S. Book Exchange) and the establishment of the Sci- entific Information Service of the Inter- American Institute of Agricultural Services for the purposes of promoting "the betterment of scientific com- munication facilities among technicians working in research, education and agricultural exten- sion." 93 A special example of an independent privately supported organization devoted to the advance- ment of libraries and documentation is the Council on Library Resources established by the Ford Foundation and incorporated in 1956 in the Dis- trict of Columbia. The Council has as its prin- "Wesseland Bejuk :, 1959 [642], p. 731. "Bush, 1945 [91], p. 107. "Murra, 1962 [402], p. 173. 19 cipal objectives: to aid in the solution of library problems; to conduct research, and develop and demonstrate new techniques and methods and to disseminate the results. It conducts its work chiefly through grants or contracts to other or- ganizations or to individuals. Over the years, then, there have been many in- stances of cooperative interplay and interaction with respect to scientific documentation among professional societies, industrial and commercial interests, nonprofit institutions, and other national organizations. Two further examples illustrate cooperative relationships in which Government agencies have also participated. First, "The Index Medicus began its career under a commercial publisher, F. Leypoldt, in 1879 .... From [1900] . . . until 1903, when Billings was able to secure a grant from the Carnegie Insti- tution of Washington, there was a gap .... With Carnegie funds the Index Medicus continued its publication based on the Army Medical Library until 1927, when it merged with the Quarterly Cumulative Index. . . . From 1927 to 1931 the Quarterly Cumulative Index Medicus was a co- operative production of the Army Medical Li- brary and The American Medical Association." 9i Second, "Making readily available the research information of the world literature in its field is the task of the Alloys of Iron Eesearch Committee which is in process of preparing monographs on the important iron-alloy systems. This useful, expensive, and still unfinished project was financed in part by Engineering Foundation, the National Bureau of Standards, and Battelle Memorial In- stitute, in large part directly by the metallurgical industries." 95 2.4. Role of International Groups When the Smithsonian Institution's first sec- retary, Joseph Henry, initiated plans for coopera- tion in international bibliographic efforts in 1818, he offered to be responsible for the indexing of the American scientific literature if the British Asso- ciation for the Advancement of Science would undertake the rest. 96 The immediate successor of this proposal, the Catalogue of Scientific Papers, and several subsequent efforts involved collabora- tion and cooperation in sponsorship, in financing, or in working contributions. Thus the Catalogue of Scientific Papers was produced from 1851 to 1925 under the auspices of the Royal Society with financial support from the Society, the British Government, individual donations and proceeds from sale of the products ; the International Cata- logue of Scientific Literature was sponsored by an international organization of participating nations in collaboration with the Royal Society; and the Concillium Bibliographicum provided biblio- graphic services on cards and an index in bulletin form financed by an international association of learned societies and individual zoologists, govern- mental grants, gifts, and proceeds from sales. 97 It is to be noted that less than a year after the founding of the American Library Association, "sixteen Americans journeyed to London to help form the Library Association of the United King- dom" and that this "1877 meeting is a landmark in international library cooperation." 98 ". . . The first international conference on bib- liography . . . convened in Brussels in 1895 . . . was not international in the sense desired because planning and issuance of invitations came too late to permit wide participation. Nevertheless an Institut International de Bibliographie sanctioned further adaptation of the Dewey system, recom- mended it for classifying bibliographies through- out the world, and authorized preparation of a Repertoire Bibliographique Universel . . . ." 99 In 1896, an international bibliographic conference was convened by the Royal Society in London. Two following conferences, in 1898 and 1900, led to the International Catalogue enterprise men- tioned above. This Conference also led to the establishment of the International Federation for Documenta- tion (Federation Internationale de Documenta- tion, or FID) which "had its beginning in the Institution International de Bibliographie, founded in Brussels in 1895 by Henri la Fountaine and Paul Otlet. In 1924, this association of in- dividuals was transformed into a federation of national and international institutions." 10 ° Similarly there was established, in 1929, with "active American participation," the International Federation of Library Associations. 101 In 1927, an International Library and Bibliographical Committee was created during the golden jubilee conference of the British Library Association ". . . The International Federation of Library Associations (IFLA) grew out of this Committee's work The chief contribution of the IFLA to the international organization of bib- liography has been to provide an international meeting ground. . . . The exchange of informa- tion on techniques, processes, devices, training, current work, and common problems is instru- mental in attempts being made to standardize forms, methods, processes, nomenclature, etc; to revise policy; and to revamp professional edu- cation." 102 "Further, both FID and IFLA, alone and in various combinations, sponsor meetings, confer- ences, seminars, and other informational pro- grams and in a number of countries on topics of timely or special interest to all types of librarians and documentalists." 103 94 Rogers and Adams, 1950 [490], p. 278. 93 Gillett, 1941 [216], p. 301. 88 Murra, 1951 [401], p. 26-27. m Ibid., p. 28-29. "a Ludington, 1954 [351], p. 194-195. ™Murra, 1951 [401], p. 34-35. ioo Murra X962 [402], p. 335. 101 Council on Library Resources, 1958 [140], p. 35. loa Murra, 1951 [401], p. 44. 103 Bonn, 1959 [68], p. 1444. 20 An example of the work of IFLA is the Inter- national Conference on Cataloguing Principles held in Paris in 1961. The aim of the Conference was to "reach agreement on basic principles gov- erning the choice and form of entry in the alpha- betical catalogue of authors and titles." 104 Library associations and other interested organizations in all countries with which contact could be estab- lished, as well as a number of international orga- nizations, were invited to participate. A "State- ment of Principles" was drafted and distributed to participants, Conference sessions were devoted to discussion and amendment of this draft. The final "Statement" arrived at received the support of a large majority of the participants. Some de- tails of the principles adopted will be covered in the section of this report dealing with the cata- loging process. 105 Further examples of cooperation and coordina- tion on an international scale are the joint IFLA- FID committees on standardization, on special libraries and information bureaus, on training and on cataloging rules; the United States National Committee for the FID, representing 23 United States private and governmental agencies plus members at large, and the European Productivity Agency sponsorship of meetings of representatives of a number of countries, from 1958 onward, look- ing toward pooling of translation efforts. Both organizations joined with representatives of ICSU (The International Council of Scientific Unions) and ISO (The International Standards Organization) in a meeting held at the Hague in September 1960 "to discuss cooperatively de- veloped programs dedicated to improving com- munication of information in the national sciences." 10G Other examples of groups and organizations concerned primarily with documentation are as follows : (a) "In February, 1952, a group of librarians and information officers constituted itself under the name of Classification Research Group and submitted to Unesco a report emphasizing in its conclusion, the importance of research into estab- lishing a standard classification scheme." 107 _(b) "The documentation organizations in the aviation field in Germany . . . collaborate in the Central Organization for Aviation Documenta- tion. . . . There is cooperation and exchange with documentation organizations in the aviation in- dustries in France and Great Britain." 108 The Institute of Information Scientists was or- ganized in London in 1958 "to promote and main- tain high standards in scientific and technical in- formation work and to establish qualifications for those engaged in the profession." 109 Among the objectives of this organization are the development and promotion of educational and training facili- ties and provision of publishing mechanisms as appropriate. Also, "a new international society in the information field has been started in Paris under the title, International Association of Docu- mentalists . . . ." 110 International bodies whose primary concerns are those of international cooperation in areas of sci- entific investigation as such include both inter- national federations of professional societies or industry-wide common interest and multi-govern- ment international organizations — agencies of The United Nations, Unesco, The Food and Agricul- ture Organization which "maintains the largest library of any international organization in the world," m multi-nation collaboration in the In- ternational Geophysical Year, and Euratom's Centre European pour le Traitement de l'lnforma- tion Scientifique, for example. Further, "the Conseil International du Batio- ment has representatives from numerous Euro- pean countries that had decided on a joint docu- mentation program. Every national member com- piles the titles of works that appeared in his coun- try in the field of structural engineering and ex- change them with all other countries. Standard form A7 (74 x 105 mm) has been selected for the international exchange. The titles are supple- mented by a table of contents in the language of publication and (on the reverse side of the card) a table of contents in English or French." 112 The International Council of Scientific Unions has already been mentioned. "The Abstracting Board is a nonprofit international organization sponsored by ICSU and incorporated in Belgium. It was established in 1950, its organization stem- ming from recommendations made at the 1949 In- ternational Abstracting Conference held in Paris. The Board's object is to achieve, through inter- national cooperation, improvement in the quality of scientific abstracting and acceleration of distri- bution of the product among scientists." 113 It is reported with respect to Unesco's programs that "in cooperation with the International Coun- cil of Scientific Unions, it will seek to extend the coverage of abstracting journals to new disci- plines, a study will be made, in cooperation with The International Federation of Documentation and other international organizations, of means of arranging for coordination between the serv- ices responsible for the translation of scientific work." 114 Perhaps under the precedents established by the League of Nations in the establishment of the In- ternational Institute of Intellectual Cooperation, 115 Unesco has been concerned with international co- 1M Murra, 1951 [401 ]i, p. 44. 105 Sec p. f.l fe. of this report. 1« O'Brien, 1963 [433], p. 203. 115 U.S. Senate. 1960 [592], p. 89-90. ,w Holmes, 1955 [209], p. 50-51. "' U.S. Senate, 19G0 [592], p. 146. ""Metcnlf et al, 1944 [380], p. 37. »»Ibid, p. 37-38. 150 Orr and Pinjjs, 1964 [441], p. 1163. "'Bralove, 1963 [75]/. p. 78. 152 Mlkhailov. 1959 [384], p. 516. 153 "List of Micropublishers", UNESCO Bull. Libs. XVI, 198-205 (1962). 49 . . . has just published the third and revised edi- tion of Photocopies from Abroad, a directory of . . . services in 38 countries. Reproduction serv- ices listed were collected during a 1962 survey with the help of national members of FID, with entries being selected according to their ability and will- ingness to serve customers abroad. More than 150 separate services are described. . . ." 154 On the subject of photocopying, Evans points out that it and the provision of microfilm and other forms of microcopying could well be insti- tuted or expanded by many reference libraries. "Only 55 of the 212 reference libraries reported a photocopying service, and in the majority of these, it is designed solely for the use of agency personnel and is not normally extended to other readers; 92 libraries indicated that they had one or more reading machines for microfilm or other microforms. Maximum use of such services can- not take place until the interests of copyright own- ers have been reconciled with those of readers." 155 Another opinion was : "Photocopying must come of age, and an organized national plan for copy- ing projects should, be established ... a report pre- pared for the American Council of Learned So- cieties . . . recommends establishment at the earliest opportunity of a national committee on the photo- copying of foreign manuscript and archival ma- terial. Other recommendations call for the estab- lishment of a national center abroad to coordinate projects engaged in copying foreign archival records and manuscripts, development of copying priorities in its field by each learned society, care- ful examination of copying proposals submitted to foundations, and a microfilm clearinghouse which among other responsibilities can insure ade- quate recording and cataloging of material copied." 156 3.5.4. Depository Libraries "We have now reviewed interlibrary loans, ex- change agreements, and photoduplication as means for cooperative acquisition of or access to scientific and technical information. Still another form of cooperative venture is the depository library, a system whereby collections of material are placed in a number of libraries, usually scattered geo- graphically but bound by ties to the central office which distributes the material. This placing and distributing can in itself be a cooperative activity, carried out by a group of libraries which main- tain a central depository for material to be shared. This latter arrangement offers a particular li- brary "a net reduction in the number of volumes it must retain for immediate access at the same time that it yields a net addition to the total re- sources of that library for deferred access. The overall gains may thus be of striking importance. The new Midwest Inter-Library Center has been planned by a group of middle western university libraries to achieve exactly these goals as well as other kinds of cooperative access to materials." 1B7 Two examples of depository libraries estab- lished by Federal agencies are those of the Atomic Energy Commission and of the Clearinghouse for Federal Scientific and Technical Information (formerly OTS). "In addition to providing information products and services to its own laboratories and those of other government agencies and their contractors, the AEC provides complete collections of AEC unclassified research and development materials to 84 domestic depository libraries and 83 deposi- tory library collections in 58 countries outside of the United States. . . . (The collections include) over 28,000 reports, some in microcard form and some in full-size copy . . . [and] a micro- reader." 158 A few years ago, ". . . On the initiative and with the support of the National Science Founda- tion," 159 OTS tried out a national depository li- brary arrangement whereby microfilm reels of De- fense, AEC, and NASA reports were deposited in 12 research libraries geographically distributed around the nation. Low usage, however, has ap- parently discouraged continuance by the Clearing- house. NASA, however, has "established an unusual search tape system that provides up-to-date, com- prehensive search capabilities at individual re- search centers of the entire NASA report collec- tion. The collection covers not only the documents announced in Scientific and Technical Aerospace Reports, but also the journal literature announced in International Aerospace Abstracts. Currently the search tapes are in use at six NASA research centers, eight principal contractors, three univer- sity centers, and at NASA's Scientific and Techni- cal Information Facility where the tapes are ini- tially prepared . . . Microfiche copies of all new reports are distributed to each organization in the search tape program. Each group is thereby en- tirely self-sufficient in that microfiche or hard copy can be supplied directly upon request." 16 ° On the international scene, there have been rec- ommendations ". . . To institute regional depots under international coordination to store complete experimental records and other documentation in support of brief published papers . . . ." 161 3.5.5. Cooperative Exchange of Data We might mention briefly a specialized form of cooperative acquisition and exchange, that of new information and data rather than of publications. Two examples will help to define this new form 164 Scientific Information Notes 5, 3, 12 (1963). "» Evans, 1963 [188], p. 28-29. See also p. 44ff and 99 of this report. 166 Scientific Information Notes 6, 4, 3 (1964). 167 Fussier, 1955 [207], p. 178. i=8 U.S. Senate, 1960 [592], p. 36, 48. "8 Green, 1963 [225], p. 224. wa Scientific Information Notes, 6, 3. 11 (1964). 161 "International Cooperation in Pure and Applied Science," 1961 [133], p. 989-990. 50 and to demonstrate why it belongs with other types of acquisition and exchange programs. The New Drug Information project (NDI)is an experiment among pharmaceutical companies (seven companies in the Fall of 1961) in the ex- change of information on biological properties of new chemical compounds, as reported in the cur- rent literature. This project was designed "to pro- vide an alerting service for the scientists of each firm that would be more comprehensive than any one company could provide without greatly in- creased costs." 162 The three military services have agreed on a pro- gram "for the interchange of data resulting from tests of ballistic missile components and parts . . . Reports of tests conducted by any ballistic missile contractor will be reported to the 'IDEP Data Distribution Center' of the cognizant service. . . . The microfilms of the complete report . . . plus a summary card . . . will automatically be distri- buted to all participating contractors and agencies which have previously expressed an interest in that particular subject." 163 It may be that these two data exchange programs forecast increasing cooperative activity in the earlier stages of research and development pro- grams. Other examples are the Industrial Liaison Office, the Cancer Chemotherapy National Service Center, and the newly established National Stand- ard Reference Data System. "The Industrial Liaison Office (ILO) was estab- lished by the U.S. Army Chemical Corps to solicit from industry data, know-how, and ideas to enhance the Corps' research and development ef- forts. Information obtained through this pro- gram has provided leads to the solution of some of the most urgent problems and has prevented duplication by the Chemical Corps of costly re- search already completed in industrial labora- tories but which has not been publicly reported. "More than 90% of the information sent to the ILO consists of proprietary lists of recently pre- pared compounds with attached physical, chem- ical, and biological screening data. At the present time, lists containing approximately 1000 struc- tures are received each month. . . . Naturally, the ILO will not release the data to other groups with- out the written approval of the company submit- ting the original sample." 1C4 "As a part of the cancer chemotherapy program, laboratries across the nation daily compile and send in test reports on drug effects on animal tumors. The data is then processed by computer, providing rapid results on the test compounds for immediate use by NIH, test drug suppliers, and the laboratories ... A master file of more than 200,000 chemical compounds and natural products screened in 26 different cancer systems has been made available to researchers." 165 102 Bohr and Owen, 1963 [67], p. 201. 163 Burnett, 1960 [88], p. 31. 164 Gelberp et al, 1962 [212], p. 7. 105 Scientific Information Notes, 6, 2 (1964). "A National Standard Reference Data System has been established by the Federal Council for Science and Technology and responsibility for its administration has been assigned to the National Bureau of Standards . . . The system will provide critically evaluated data in the physical sciences on a national basis, centralizing a large part of the present data-compiling activities of a number of Government agencies . . . ." 166 3.6. Analysis and Identification After the existence and availability of docu- ments have been determined and they have been acquired by some means or other, they must be processed before being put away for later refer- ence. This processing consists of anatysis of the form and content of the documents, plus identifi- cation and recording of the results of such analysis. We shall discuss these various aspects in turn : physical identification, content identification, and cataloging rules and principles devised to aid these identification operations. We shall also examine current operations and proposed activities for cooperative or centralized cataloging and for the use of automatic equipment in conventional processing steps and in the newer approaches to cataloging and indexing. 3.6.1. Physical Identification — Descriptive Cataloging The physical description of a document is the most basic of the operations taking place in a li- brary or information center; before the organiza- tion can attempt to handle a document, it must know what it has. In this section we will discuss the more important elements of a descriptive cata- loging entry, particularly as applied to technical reports. In the section on cataloging rules, (3.6.4) early efforts in the development of com- mon practices among libraries will be reviewed, particularly the origin and growth of standards for a library catalog. The problems and difficulties of descriptive cata- loging assume special significance in terms of fu- ture mechanization. For one reason, the rules involve so many complicated alternatives that they have to date defied the programmer's ingenuity. For another, it is precisely the card catalog opera- tion that has received the greatest attention with respect to the possible introduction of machine handling in very large systems. Descriptive cataloging is defined by Webster's New International Dictionary, Third Edition, as "a library procedure by which a book or other item is identified and described by recording such items as author, title, imprint, and collation — contrasted with subject cataloging." Taube comments that descriptive cataloging is not the "relatively simple and straight-forward affair" that it might seem, that there is difficulty in "devising uniform entries from the haphazard inf ormation which appears on 1OT Scientific Information Notes 5, 4, 1 (1963). 51 the title-pages of the various publications and re- ports requiring organization." 167 Although, in the opinion of one author, "It is understood that descriptive cataloging is stand- ardized. Consequently, there should be little difficulty in creating a standard for descriptive cataloging," 168 many workers in the field would not agree. For example, "the prevailing rules of cataloging are based on the Anglo-American rules of 1908. These have frequently been amended since that time, however, and as a result have become so com- plicated and difficult of interpretation that the value of the code has in good part been diminished . . . ." 169 Further, "one of the favorite demonstrations of those who run the National Union Catalog at the Library of Congress is to hold up 10 catalog cards from the major libraries of this country, all repre- senting the same book, and all beginning with a different main entry ; and this despite the fact that all 10 of these libraries were attempting to inter- pret the same set of rules !" 17 ° Much discussion and even argument has gone on and continues about rules for descriptive catalog- ing. It could be a most fruitful area of coopera- tion and collaboration, but agreement is difficult to attain. We might add here that many workers in automatic system design claim that descriptive cataloging is more difficult for a computer to ac- complish than automatic indexing or abstracting. The difficulties will become apparent as we discuss the individual problems. The traditional means for bibliographic control of a book or a publication has been citation of the author. Items within periodicals are generally referred to by the personal authors of those items. Even those libraries that do not have a common dictionary catalog containing an alphabetical ar- rangement of cards indicating authorship and sub- ject matter of publications usually have separate catalogs for authors and for subjects. With the advent of the technical report in this country, the problems of identifying the organiza- tion sponsoring the research or the investigation assumed increased importance, especially where no personal author was identified or where many different authors prepared various reports on the same projects. In the AEC, for example, use of the personal author entry led to some confusion "because the reports were produced by a team of research workers . . . When work is performed in this manner, it is not unusual for the first paper to be prepared by one or two members of the team, the second by a different pair and the third by the team as a whole." 171 Uniform identification of corporate source has been a perennial problem, in view of frequent changes of name, which may require the mainte- nance of separate history files (Mac Donald, 1953 [364] ) . But for the technical report the organi- zation must often be regarded as the main entity for entry into the catalog. This involves the use of the corporate author for announcement, cata- loging, and filing purposes. The personal author is considered of secondary importance. "Although personal authorship is important, the actual iden- tification of a report must be tied to the corporate agency responsible for the work." 172 Current methods used for designation of cor- porate authors are complex and difficult to fol- low. This is partly because ". . . Laboratories place their names on a given series of reports with considerable inconsistency and . . . many labora- tory names are homonymous and undistinc- tive." 173 One solution to the complex problem was offered by Taube, who suggested that wherever possible "the most specific corporate body involved was to be employed. Treatment of the most specific body as a subdivision of its parent body was to occur only when it had a name which was capable, and likely, of being a subdivision under many other corporate bodies. Names falling into this category were not defined in a general way, but were speci- fically promulgated in a list." 174 Another approach to solution was undertaken in 1951 when the Board on Cataloging Policy and Research of the ALA Division of Cataloging and Classification "initiated a study of the second edi- tion of the Code by Seymour Lubetzky, with par- ticular attention to the rules for corporate authors and a discussion of the objectives and principles which should underlie a possible revision. This study was published in 1953 as Cataloging Rules and Principles . . . "Mr. Lubetzky pointed out a number of in- stances of poor arrangement of rules, cases of in- consistent treatment of similar conditions in dif- ferent sections of the code, and examples of unnecessary repetition of rules. He gave particu- lar attention to the rules for corporate headings and showed how these had developed into the pres- ent confused sections on societies, institutions, and miscellaneous bodies." 175 The complexity of the methods used for desig- nation of corporate author is shown by the rules devised by Lubetzky in an unfinished draft for a new edition of cataloging rules prepared for the ALA Cataloging Code Revision Committee (1960 [349] ) . He divides the works of corporate author- ship into the following types and gives a rule for establishing the entry in each instance: work of corporate body; work of division of corporate body; work of group organized or sponsored by corporate body ; work of unnamed group ; work of individual that is issued by a corporate body. "'Taube, 1950 [556], p. 3. 168 Lanjrenbeek, 1962 [3291, p. 298. 169 Scientific Information Notes 4, 5, 3 (1962). 170 Brandhorst. 1964 r76], p. 43. I71 Croxton, 1955 [151], p. 126. 172 Warheit, 1952 [627], p. 105. 173 Brandhorst, 1964 [76], p. 42. w« Ibid., p. 37. 175 Wright, 1956 [655], p. 331. 52 Government publications may have the following additional rules : work of national or local juris- diction; work of jurisdiction subject to authority of another jurisdiction; work of government de- partment, office or agency ; work of government officials. The DDC has compiled and revised its Cor- porate Author List, the most recent list being is- sued in April 1963. The ASTIA Guidelines for Cataloging and Abstracting establish the policy of the Center regarding the List stating that "the corporate author (corporate body directly respon- sible for preparing a document) is treated as the main entry for each accession cataloged into the DDC collection, and except for the accession num- ber, represents the principal approach to individ- ual holdings." The AEC also has a list of corporate author en- tries. The list currently available is Corporate Author Entries Used by the Division of Technical Information in Cataloging Reports, publication number TID-5059 (5th Rev.), issued in March 1962. It has an appendix listing the rules for cor- porate entry of the AEC, the main principle being : "Scientific and technical reports are to be entered under a corporate form of entry in addition to entry under the personal names of the authors. The corporate entry is to be based on the name of the issuing body current at the time of publication which is experimentally (or technically), editori- ally and/or contractually responsible for the re- ports." In more detail, the AEC, instead of treating all corporate bodies alike, defines several categories of organizations (eight, to be exact) and makes the old distinction between societies and institutions, developing different rules for each category. In addition, AEC demands research beyond the docu- ment in hand, so that Ladd Observatory is entered Brown Univ., Providence, Ladd Observatory, but Lick Observatory is entered Lick Observatory, Aft. Hamilton, Calif., rather than under its parent, the University of California. 170 NASA has made several listings of its corporate authors, with codes assigned to each one. Such lists are apparently not intended for widespread distribution, probably because of the belief that there is no need for them outside NASA. A group of government representatives, com- prising a modern-day group for the standardiza- tion of information services and calling itself the Joint Descriptive Cataloging Group, held meet- ings in April, June, and September 1963. It pro- posed to advise the Committee on Scientific Infor- mation (now COSATI) of its recommendations as to procedures which could lead to standardiza- tion of descriptive cataloging by the producers and distributors of government research reports. A Standard for Descriptive Cataloging of Technical Reports (1963 [195]) was published in December 1963 by COSI. The rules for corporate author 170 See Brandhorst, 1964 [76], p. 39. entries in the Standard were included "in the hope that fruitful discussion of the degree of specificity will be obtained regardless of agreement on the order in which the chosen heading is written. It is planned that a working group representing con- cerned agencies will undertake the preparation of a combined corporate author heading list." A special case with respect to main entry cata- loging and an extension of the problems of proper determination of corporate authorship has been raised by Gull. He proposes that rules be drafted for entry and descriptive cataloging for auto- matic authorship, or nonhuman authorship. He points out that "electronic and mechanical de- vices are already producing mathematical and technical texts (excluding belles-lettres) to which cataloguers can attribute neither personal nor corporate nor anonymous (i.e., personal but un- identified) authorship", and contends that rules must be developed to care for "all works and rec- ords of personal, corporate and anonymous author- ship susceptible to electronic and automatic equip- ment," 177 The sum of all these problems and suggested solutions to the author citation question point to ". . . the eventual need for a Directory of U.S. Corporate Authors in the Scientific and Technical Area ... It would serve as a massive influence toward standardizing cataloging practices and reducing cataloging costs throughout the special library and science-technology information center field. It would be an instrument by which the relationship between existing practices could be displayed to facilitate choice, conversion, etc." 178 Even if the questions of content and order within author and corporate author entries can be resolved, an additional problem remains as to the form of the entry, whether it follows the tradi- tional library practice of main organization with subdivisions following or whether it uses the smallest significant component. There are draw- backs to both forms, but either system can be made to work "by the generous and judicious use of ref- erences. It really makes no major difference which form is used as long as you decide on one and stick to it. In either case an authority book or listing of the names accepted is essential." 179 Many efforts have been made to formulate rules and standards for descriptive cataloging. For 50 years every international library conference has expressed interest in bringing about agreement on such rules. One of the major roadblocks to ex- tensive agreement has always been the conflict between two major cataloging traditions in the matter of corporate authorship. The European tradition, taking its lead from the so-called Prus- sian Instructions, does not recognize the principle of corporate authorship, and prefers to enter works under their title or personal author. The 117 Gull. 1963 [230], p. 2S4. ™ Brandhorst. 1964 [76], p. 43. > 7 »Croxton, 1955 [151], p. 127. Do Anglo-American tradition, "branching from Pa- nizzi, via Jewett and Cutter," champions the prin- ciple of corporate authorship but finds it difficult to apply in specific situations. It is necessary that the rules now used for this aspect of cataloging be substantially simplified. "They must be shorter, easier to understand, quicker to use, and capable of greater consistency of application; and their re- sults must be better directed towards the legitimate functions of the catalog and the needs of the users, and less costly in their operations." 180 Another entry in descriptive cataloging of an item is the title. The informal nature of report literature sometimes makes the selection of a title difficult, since some reports have multiple title pages, some have cover titles, and some show no titles. It is essential that the information given as the title be informative as well as accurate, and that the substance of it be quickly understandable. To this end, titles beginning with such words as Progress Report are often inverted so that the subject information appears first. 181 The determination of the date of a report is an- other problem because many reports are issued either without dates or with multiple dates being given. "The date of a report is of considerable importance to the user in judging its worth and in determining the priority of research effort." 182 The date is also important in cataloging for iden- tification of delayed proceedings, collections of hitherto unpublished papers and talks, and manu- facturers' brochures and manuals. Still another area of descriptive cataloging which receives continuing attention and effort to- ward standardization is that of citing journal ref- erences, especially in abbreviated form. One tool aimed at helping this effort is the Coden for Pe- riodical Titles, prepared and maintained for Com- mittee E-13 on Absorption Spectroscopy of the American Society for Testing Materials (Kuent- zel, 1963 [324] ) . The book contains nearly 20,000 coden for titles. The term "coden" is defined as "the combination of letters, numbers, and symbols assigned to a document or other item as the result of applying a set of coding rules." 183 The coden are sets of presumably unique four-alphanumeric character codes for the titles of periodicals. How- ever, many possibilities of homographic ambigui- ties are evident and obviously involve a continuing problem area. The ASTM list of coden was derived from one desecribed by Bishop in 1953 [62]. The most com- monly used format gives a title code first, followed by volume and page numbers and year of issue. Thus JACS-77-2282-55 indicates a reference to page 2282 of volume 77 of the Journal of the Amer- ican Chemical Society published in 1955. The coden use four-letter codes only for periodicals; nonperiodicals, patents, and one-time publications 180 Brandhorst, 1964 [T6].. p. 36. 181 See Croxton, 1955 [151], p. 128. 183 Ibid., p. 129. 183 Bishop, 1953 [62], p. 58. receive a different code. While the coden for peri- odicals are mnemonic, those for proceedings of a symposium, minutes of a meeting, or a collection of papers consist of two digits and two letters, bearing no mnemonic or other relation to the pub- lication but serving to identify it via an index. Other types of abbreviations for titles of peri- odicals have been used for years in the compila- tion of bibliographies and "lists of references," and there have been many different rules devel- oped to abbreviate the titles. Impetus should be given to the attainment of standardization of the abbreviations by approval of the American Stand- ards Association American Standard for Periodi- cal Title Abbreviations. The standard was prepared by a subcommittee of the ASA Section Committee Z39 on Standardization in the Field of Library Work and Documentation. The Subcom- mittee on Periodical Title Abbreviations was or- ganized in December 1961 and began compilation, with the use of punched card equipment, of a pri- mary list of periodical title words that are fre- quently abbreviated. This primary list was com- piled from lists of abbreviations supplied by Index Medicus, Biological Abstracts, Bibliog- raphy of Agriculture, and Chemical Abstracts. The standard gives definitions of terms and rules for omission of letters, word order, single word titles, articles, conjunctions and prepositions, capitalization, punctuation, plurals, compound words, clarification in abbreviation, and multi- lingual abbreviations. After the composition of the citation has been stabilized, its form must be considered. The GSIS card with the citation to the left and main entry by corporate author and tracings to the right "has the advantage of being applicable in the direct photographic preparation of bibliographies using cards from any or all . . . cooperating agencies." 184 The question of photoreproduction raises the further example of complications in descriptive cataloging : "The catalogue entry of a publication in microform should be identical with that of the original. The first supplementary note should in- dicate the microform and whether it is a positive or negative copy . . . For positive copies the loca- tion of the negative should be indicated, if it is known." 185 3.6.2. Content Identification— Subject Cataloging The Utopian ideal in literature indexing, uni- versal bibliography or indexing, means indexing everything in a piece of literature, "intentional, incidental and accidental, that could possibly be of interest to anyone, anywhere, anytime." 186 This is not practical even with full text search "by some mechanical brain indexing everything on everything." 187 Instead the operations of subject cataloging attempt to provide at least some meas- "« Croxton, 1955 [151], p. 131, see also p. 14-15 of this report. 185 Gunther, 1962 [232], p. 10. 188 Metcalfe, 1959 [381], p. 12-13. 181 Ibid. 54 ure of control, to indicate what a book or docu- ment "is about", and to provide a central means of access to all items in the collection which are "about" some subject or subjects. For example, in a printed index or card catalog, the objective is the entry of all that is indicated on a given subject at one point in the catalog. Single volumes and sets of volumes, constituting bibliographically complete works, are cataloged by their general subject or subjects; in addition some parts are recognized as linked bibliographical units, such as chapters and essays in books and articles in periodicals, and theses may be indexed for their general subject contents. This indexing or analysis is fairly completely and systematically done for some subjects or some periodicals in peri- odical indexes and abstracting journals such as the British Library Association's Subject Index to Periodicals, the Wilson Indexes, Plant Breeding Abstracts, Chemical Abstracts, and many others. Subject cataloging is not however as easily and consistently accomplished as might seem possible at first glance. Efforts are made to develop rules and principles for subject cataloging as for de- scriptive cataloging, but the task is long and diffi- cult. Thus, serious efforts will continue to be made in order to make subject cataloging easier, more consistent, and more useful for the searcher of a collection. Toward these ends, many tech- niques have been proposed or are in practice for the analysis and recording of the subject content of documents. Systems of documentation may be used in the selection, dissemination, and retrieval of informa- tion (or more properly of "documents containing information"). Such systems for retrieval of in- formation "have been created because it was im- practical, inconvenient, or excessively expensive to locate records personally." 188 Retrieval has been accomplished traditionally by schemes for index- ing and schemes for classifying. These schemes assist a searcher, inspecting a file or library of records, to find particular records usually in terms of their subject interest. Such a searcher must (1) predict the vocabulary of authors, probably unknown to him, in dealing with matters which in- terest him; (2) predict the terminology with which librarians or document alists, probably unknown to him, have analyzed the records; and (3) predict the symbols or other clues used by these librarians or documentalists to record the results of their analyses. Indexing is the usual prerequisite for both the selection and retrieval of documents and the infor- mation in them and for selective dissemination of documents. It creates or uses one or more previ- ously created surrogates for the content of a docu- ment and indicates or records those surrogates associated with the document. An index is ". . . 'anything used to indicate, point out, or guide.' The process of indexing involves selection of words or ideas from a graphic record on the basis of well- defined rules ; the purpose of indexing is to facili- tate the identification or selection of desired graphic records (documents, books, etc.) after they have been sorted and shelved or stored." 189 In traditional American librarianship, the pro- vision of a single point of access to items having similar subject content has been by means of the selection of an appropriate term or terms from a set of alphabetically arranged subject headings. A subject heading is defined by Frarey as a "word or verbal expression deliberately chosen from among the various alternatives to express the particular content of the material which it de- scribes and which will be (1) in harmony with the usage of the audience to whom it is addressed, (2) accurate and precise in its specification of the exact subject of the material, (3) uniform, in the sense that the same word or verbal expression will be used consistently to describe the same subject, and (4) amenable to integration with other subject headings to provide a usable arrangement. In American library practice . . . the use of subject headings implies an alphabetical arrangement." 19 ° The sixth edition of the Subject Headings Used in the Dictionary Catalogs of the Library of Con- gress states : "The present edition, like its predeces- sors, is a record of the headings traced on the Library's printed catalog card and used in its card catalogs and cumulative Catalog series. The list is the product of evolutionary forces, among them the growth of the Library's collections, semantic changes, and varying theories of subject heading practice over the years. As a consequence the list is, at any point in time, an accurate reflection of practice but not a complete embodiment of theory." 191 In the summary to his review and evaluation of literature concerning the alphabetic subject cata- log, Frarey makes 11 points, among which the fol- lowing are pertinent : ( 1) Present theory and prac- tice of subject cataloging "is based primarily upon tradition and assumption and does not reflect any clear understanding of function or purpose." (3) Evidence indicates that the subject catalog "will be found to serve best the needs of any user for general purposes, i.e., it will orient him to the subject area of his inquiry; it will help him to select the best material available for securing this orientation ; and it will guide him to the other in- struments for subject access in his field of inquiry if his needs extend beyond those of simple orienta- tion." (8) Those imperfections and inconsisten- cies in present practice which have been shown by studies of terminology, form and structure and which cannot be eliminated or minimized "will have to be resolved by the adoption of standards ""Kent, 1063 [311 J, p. 2G7. "o Ibid, p. 2C.7-268. 100 Frarev, 1»60 [202]., p. 5-fi. 101 Library of Congress, 1957 [338], p. ill. 55 based upon sound knowledge of linguistics, seman- tics, and grammar, study of which in relation to subject catalog terminology and structure is only in its infancy at the present time." Further, (9) the syndetic reference structure (provision of see and see also references) is infrequently used and has even occasionally been abandoned altogether, thus implying that the provision of a comprehen- sive inventory by a subject catalog does not derive from an actual or a significant need, and that, in any case, the ability of the reference structure to do this is limited by the practical limitations which have been imposed upon its development. (11) The ability of the subject catalog to give some measure of satisfaction emphasizes "the need served by it and that a subject catalog is likely to continue to be an essential feature in effective library service. The evidence suggests further that there is an excellent chance that a highly effec- tive subject catalog can be effected." 192 Becker and Hayes (1963 [48]) regard the cata- log of the library as its most important tool for management of its collection. They believe that the artificial or arbitrary decisions required of an indexer at the input stage when he makes his initial assignments to documents are a drawback to the subject heading system, although the question can logically be asked whether this is not the case in any human indexing scheme. The semantics of a heading change with time, as do the concepts underlying an indexer's choice. The subject head- ing system also has the disadvantage of increasing rapidly in terms of size and complexity. Morris, in writing about the duality concept in subject analysis, was concerned that the tenet "voiced most persistently today is that the more 'direct' and 'specific' the heading the better the subject analysis." 193 Such headings are illustrated by the alphabetically arranged sequence: Carbon steels, Extra high carbon steel, High carbon steels, Steels. Indirect entry (by qualifying inversion) is made to the principal noun common to all four headings which then becomes the guiding word for their alphabetic sequence in the Catalog: Steels, Steels (Carbon), Steels (Carbon, extra high) , and Steels (Carbon, high) . Morris was convinced "that in modern subject cataloging, the phrase 'direct and specific' is syn- onymous with 'not subdivided extensively and not inverted.' The most that can be said for this as a basic principle is that in practice it succeeds in separating (albeit scattering throughout the al- phabet) nonidentical data but it does practically nothing to the end that similar or related data will be brought together. Carried to its full implica- tion it renounces completely one of the fundamen- tal purposes of subject cataloging, the twin of the duality concept concerned with giving the user some compresensive view of sizable segments of a collection and bringing pertinent relationships into juxtaposition through a card catalog. This becomes doubly significant if, for large areas of the library collection, subject cataloging must carry also the burden of substituting for library classification of data." Wi The other twin of the duality concept of Morris is "to enable a user of a library catalog to determine readily whether the library has available the particular bit of infor- mation which he desires . . . ." 195 Concept coordination, or coordinate indexing, another approach to information storage and re- trieval, "was discussed in theory in the latter part of the 19th century but was applied only in several isolated situations until the beginning of World War II. In the years since the war, concept co- ordination has gained progressively more accept- ance as inherent problems have been solved which originally limited utility and effectiveness. . . . In concept coordination, information is analyzed and characterized for storage and subsequent re- trieval primarily by single word units to describe single ideas or unit concepts." 196 The striking characteristic of coordinate in- dexing systems is that the single terms used for in- dexing are not combined in advance for storage, as is true with the main headings plus modifiers of subject heading systems. Coordinate indexing systems are said to be post-coordinated; that is, combinations of the terms which describe a ques- tion are made at the time of asking the question. Further, "the basic principle of 'coordinate index- ing' is that of free combination of concepts ... to define a multidimensional search . . ." 197 Concept coordination systems are found in sev- eral forms, depending on the amount of complex- ity or structure present. Some systems even con- duct some amount of precoordination. Coordinate indexing is defined more broadly "to include all systems in which the logical operation of intersection, union, and negation are brought into play in the manipulation of index terms. . . . whether Uniterms, keywords, descriptors, unit concepts or structerms. . . ." 198 Examples will illustrate the different types of systems. The Uniterm system of coordinate indexing, de- veloped by Taube in the early 1950's, described the informational content of documents, "after appro- priate analysis by qualified personnel, by the sim- plest practical w T ord units of information referred to as Uniterms. Uniterms are subject headings for unit concepts, with all Uniterms possessing equal hierarchical stature, none subordinated or superordinated to any other (as in classification) and none existing in precoordinated arrangement with any other (as in traditional subject heading arrangement). Taube visualized Uniterms as predominantly single words." 199 i^Frarey, I960 [202], p. 63-65. lea Morris, 1954 [395], p. 121. 194 Ibid, p. 122-123. 195 Ibid. p. 117. 199 Costello, 1961 [138], p. 21. 197 Taube et al. 1954 [562], p. 62. 198 Jaster. 1962 [296], p. 23. I99 Costello, 1961 [138], p. 21. 56 "In the summer of 1952, the Armed Services Technical Information Agency awarded a con- tract to Documentation Incorporated for an in- vestigation and experimental installation of a then completely novel system for filing and retrieval of information, the Uniterm System of Coordinate Indexing . . . "Coordinate indexing as a generic term covers all forms of indexing in which the retrieval of specific items of information involves the determi- nation of the logical product of a number of classes . . . "The Uniterm System as a species of coordinate indexing is a manual method of determining the logical product of two or more classes through the device of 'arithmetical' coordination. The dis- covery of a common number on two or more Uni- term cards establishes that there is a class which is the logical product of the classes denoted by the Uniterms and that the class has members. The members are, or course, the documents or other items designated by the common numbers." 20 ° Changes in Taube's original system between 1951 and 1961 affirmed the importance of the prin- ciple of bibliographical coordination even though they diverged sharply from the simplified initial concept. Among the changes which have been made in systems originally based on Uniterm are rejection of equal values for terms, introduction of hierarchial relationships; retention of subject authority and cross reference systems, and the in- troduction of "role indicators" to serve as standard subdivisions or modifiers and "links" to prevent undesirable coordinations among terms. Artandi and Hines argue that the changes introduced by users of coordinate indexing "show, although this is obscured by the terminology, a steady growth of acceptance of the use and utility of major elements of conventional subject-heading practice as it has developed from Cutter's basic work in 1876." 201 Premodification or precoordination of terms (the compounding of a plurality of terms substi- tuted for single terms when necessary to specify the subject accurately for the purpose and search method at hand, e.g., Boundary layer or Aircraft -fire control systems) resulted in more or less highly controlled terms. Another method of con- trolling the terms for a concept coordination sys- tem is by using "descriptors," general terms whose meanings encompass meanings of several other terms ; descriptors were devised to reduce the num- ber of terms in coordinate indexing vocabularies. By definition, a descriptor system has a limited vocabulary; a descriptor dictionary is usually maintained, with "scope notes" to define the scope and meaning of each descriptor for the system. The Zatocoding System of Mooers uses descriptors in this way (1951 [394]). Mooers has described his descriptors as follows : "The neologism 'descriptor' was purposely intro- duced in 1950 in order clearly to distinguish the new method of indexing language from the other and older systems such as : word indexing, stich- wort classification, schlagwort, subject heading, etc. . . . "We may think of the whole descriptor, includ- ing the definition, or scope note, to which the label is attached, as a packaged idea, as a concept which is to be used primarily for retrieval in some par- ticular retrieval system. . . . "Another characteristic, which is rather subtle, is that the set of descriptors as a whole, and the descriptors individually, are tailored to do a par- ticular retrieval job." 203 More complex and structured schemes for con- cept coordination systems use role indicators and links to modify the basic terms and to prevent "false drops" or inaccurate coordinations. The highly complex abstracts of documents or "telegraphic abstracts" devised by Perry, Kent, and Berry are another effort at reducing the prob- lems of false drops and the difficulties with subject interrelationships occurring in the coordinated index or superimposed coding types of schemes. According to Vickery, the central semantic problem of subject indexing is the relation between terms, and the three following modes of analysis, distinguished in traditional logic, are used in dif- ferent systems of information retrieval : ". . . the physical analysis of a thing into its parts or con- stituents, or of a group into its members; the logical analysis of a generic concept into its spe- cies; and the metaphysical analysis of a concept into its attributes. A special instance of the last is the analysis of the definition of a concept into its elements." 204 Vickery sees also three tech- niques of semantic analysis in selection of index- ing terms: the "analysis by definition" of Perry, Kent, and Berry (1956 [455]) the "analysis by operational definition" of Andrews and Newman (1956 [31] ), and the "facet analysis" of Bangana- than (1957 [469]). Perry et al., have analyzed words in science and technology and find that "semantic factors," i.e., terms not further analyzed, and combinations of these factors are sufficient for representation of the words. Andrews and Newman represent a word by the combination of a limited number of attri- butes, providing a series of "modulants," of cate- gories or inflecting codes, which serve to form descriptors by inflecting the "ruly roots" for things named in the literature (the "ruly" roots being opposed to common "unruly" parlance). In the facet analysis of Banganathan, each of the terms in a subject field is defined with respect to its parent class. The terms are then sorted into the categories formed, such as substance, state, prop- erty, reaction, operation, or device, so that these categories can be combined to form compound terms. 200 Tnube et al, 1954, T562], p. 44^16. 201 Artandi and Hines, 1963 [35], p. 77. 203 Mooers, 1963 [392]], p. 27. =" Vickery, 1959 [613]i, p. 855-858. 57 In these three types of analysis the choice of semantic level of indexing terms is aided or con- trolled by the formulation of categories which are concepts mentally fabricated with reference to the world of experience. "There is general agree- ment that the most helpful form of classification scheme for information retrieval is one which groups terms into well-defined categories, which can be arranged in hierarchies where this conforms to the recognized structure of relations between them." 205 A quite different approach to control of vocabu- lary and semantic and syntactic relations of terms has been described by Walton of the David Taylor Model Basin, Navy Department, who has devel- oped an artificial procedure-oriented language FROLIC (Formal Retrieval-Oriented Zanguage for /ndexing Content) , by which a machine may manipulate those multifaceted descriptions of sub- ject content of documents generated by analytic- synthetic techniques. The language "is built around a categorized and hierarchically arranged vocabulary combined with a simple grammar for indicating certain essential roles or relationships between index terms. . . . The key to the proposed system is a thesaurus-dictionary in which word associations and meanings are recorded. . . . The basic or elemental concepts of the language are displayed in a Classified Schedule, where they have been gathered into a few general categories. The primary division is into the names of things and the various characteristics that can be ascribed to them, viz, properties, conditions, actions and relations.'''' 20S The division, things, is classified into physical substances and objects, the former being divided into energy and matter and the latter into natural objects, bodies (masses), and constructs (artificial ob j ects ) , each of which is further divided. Walton suggests that "the classification of basic concepts (rather than subject headings), along with seman- tic analysis of complex concepts, should help to unmask much current technical jargon, while the development of an interdisciplinary vocabulary of wide application in scientific documentation should ultimately lead to greater compatibility between systems." 207 Differences between subject headings and the terms, simple or structured, for post-coordination systems are useful for categorizing indexing tech- niques. A few other distinctions of categorization merit attention. The steps to be taken by an analyst in the selec- tion of indexing terms to represent a document may be outlined as follows: "The text is scanned to select a set of words, phrases or sentences which collectively represent its subject. ... a decision is taken as to which of these subject descriptions are worth recording as being relevant to the purpose 205 International Federation for Documentation, 1957 [2851], p. Hi. 206 Walton, 1963 [626], p. 22. 207 Walton, 1963 [626], p. 22. of the retrieval system. . . . the relevant subject descriptions are transformed into the standard de- scriptor language used in the system." 208 In scan- ning the text two methods are commonly em- ployed : "First, the indexer reads the text and can 'understand' it — i.e., he can formulate the subject in his own words, which need not be the words used by the author. Second, the indexer cannot 'under- stand' the text in this sense, but he picks out from it words, phrases or sentences which the author has emphasized as important — the title, introduction, section headings, conclusions, summary, and so on » 209 This distinction between methods is reflected in contrasting "derivative" with "assignment" index- ing (Luhn, 1962 [353] ; Herner, 1962 [255]). By this method of differentiation, "derived" terms are taken from the texts of the documents themselves while "assigned" terms are taken from a list of terms or subject categories that exists independ- ently of the documents. Indexing by assignment is obviously more likely to be concept indexing than word indexing, and indexing by derivation is word indexing rather than concept indexing. Herner believes that the Uniterms of Taube, at least initially, were derived and that "Keyword indexing and/or permutation indexing, whether based on titles, amended titles, or texts, constitutes an extreme example of indexing by derivation." 210 As examples of indexing by assignment in non- conventional systems, Herner cites both Zatocoding and the work of Schultz in the man-computer indexing of the 1960 Federation of American Societies of Experimental Biology (FASEB) papers. Herner states that elemental to Zatocod- ing "is the analysis of documents in terms of care- fully developed sets or classes of criteria that guide the use of and application of substituent indexing terms or 'descriptors.' . . . Zatocoding tells you when and how to select words or descriptors in a vocabulary. In so doing, it helps to diminish re- dundancy and lack of consistency and con- text. . . ." 211 In the work of Schultz with the FASEB, "authors of papers selected appropriate index entries from a list of possible entries fur- nished them by the Federation and also from the terms used in the abstracts of their papers." 212 Kent makes a similar distinction between word indexing and controlled indexing. Controlled in- dexing "implies a careful selection of terminology for storage in the index, in order to avoid as far as possible, the scattering of related subjects under different headings. The control may be imposed by limiting the indexing (a) in the number of sub- jects that may be chosen, (b) in the number of aspects of subjects that may be chosen, and (c) in the language used to express the results of the analysis." 213 208 Vickery, 1961 [6I16J, p. 15-16. 208 Ibid, p. 16. 210 Herner, 1962 [255], p. 5. 211 Herner, 1962 [255],, p. 5. 212 Schultz, 1963 [509], p. 9. 233 Kent, 1963 [311], p. 268. 58 Four "schools" of documentation with respect to the selection of indexing terms are discussed by Becker and Hayes ( 1963 [48] ) . These schools are of (1) Taube of Documentation, Inc., Bethesda, Md., with coordination of concepts by use of Uni- terms, usually in the form of single words or con- cepts; (2) Mooers of Rockland Research Institute, Inc., Cambridge, Mass., who uses "descriptors" and "superimposed coding" for efficient storage by compacting codes randomly in a limited storage area; (3) Perry and Kent, formerly of the Center for Documentation and Communications Research of Western Reserve University, Cleveland, who in- troduced "role indicators," "semantic factors," and "analytic interfiles," and also the "telegraphic ab- stract" based on elements for describing a docu- ment's information content; and (4) the facet analysis and colon classification schemes of the school of S. R. Ranganathan, which postulates that the relationships among things in a universe can be described in terms of their positions with respect to each of a set of fundamental facets. Maloney and Batchelor relate indexing terms to words in languages and to the types of languages evolved during the last century. One of their tentative principles of information retrieval is that "retrieval index entries take three forms of isolating, agglutinative, and aggregative codes that words in languages do. Isolating codes show a minimum of 'structure' and the greatest inde- pendence of context and order in the specification of meaning. Descriptors, uniterms, and key words are examples of such codes. Most classifications are examples of agglutinative codes, but colon, UDC, and other faceted, classifications, as well as semantic factor codes, are examples of aggregative codes." 214 There have been perennial arguments pro and con "word" or "derived" indexing and "assigned," "controlled," or "subject" indexing. Crane and Bernier say that word indexes for articles in pe- riodicals that are merely indexes of titles of papers or of abstracts are to be regarded as generally incomplete: the titles frequently do not reveal, even in a broad sense, the contents of papers. "Many so-called subject indexes are really indexes of words instead of subjects. There is a vast difference. Word indexing leads to omissions, scattering, and unnecessary entries." 215 An index entry as used in Chemical Abstracts includes a heading (word or phase selected to act as guide to the subject, concept, author name, etc.) , a modification if necessary (additional statements explanatory of the individual heading), and a reference (for finding the item from which the index entry was derived). "The subject index is designed to be a key to the information, not to be the information itself. Index-heading terms usually represent more or less general areas of in- n * Maloney and Batchelor, 1963 T369], p. 108. 215 Crane and Bernier, 1958 [147];, p. 515. formation in the document, but do not give spe- cifically the information itself." 216 Welt (1959 [640] ) also believes that word index- ing (as opposed to subject indexing) is of dubious value for indexing periodicals. Subject indexing, however, involves rigid standardization of terms. The lack of standardized subject headings and their flexibility or inconstancy both lead to a sig- nificant number of failures to retrieve the infor- mation that is desired. Holmstrom suggests: "In the lexicography of science it is of first importance to appreciate and maintain the distinction between concepts — the ideas or notions, phenomena or applications which science uncovers, examines and exploits— and the terms which scientists, in their reasoning and communications, use as symbols to denote these concepts." 217 He continues that "the concepts, once established, are fixed and immutable, but the terms which scientific writers use as symbols for them are fluid and susceptible to change. To a considerable extent they depend on nothing more solid than fashion." 21S Furthermore, except for Chinese and some Japanese "the terms are symbols not for the concepts directly but for the sounds which those who happen to have been born in a particular country are accustomed to utter or to imagine when they have particular concepts in mind." 219 Holmstrom stresses that the concept of an action or a thing is not always known by the same word or words in a given language. It is for such reasons that the different types of indexing, derived and assigned, precoordinated and postcoordinated, tend to converge. The prob- lems of construction and use of thesauri, to be dis- cussed in section 3.7.1, are not very different from the traditional work of establishing, standardizing, and revising subject heading authority lists. In this area, we note a particular example of proposed cooperation on a broad scale : in 1954, Angell pro- posed that a national committee be formed on the initiative of the Division of Cataloging and Class- ification of ALA, and that it should be a "joint committee on standardization of subject headings. There should be representatives of such national associations as the ADI, the Association of Col- lege and Reference Libraries (particularly the Reference Libraries Section), the Association of Research Libraries, and SLA; of such maintainers of information systems as the H. W. Wilson Com- pany and the LC; and of the informal group of Government agencies working on these problems under the title Group for the Standardization of Information Services . . . [The committee] should formulate the objectives and principles, identify the research studies and form them into an integrated program, lay out the steps of proce- a» Bernier and Crane 1962 [551', p. 2" Holmstrom, 1955 [270], p. 73. 218 Ibid, p. 74. 2,9 Ibid. 117. 59 dure precisely, set up a budget, get the money — and supervise the execution of the work." 22 ° This probably will have to be done to standardize the universality of language to be used for storing information among various groups in this country. In summary, we might quote Herner to the effect that "the phenomenon that best characterizes mod- ern documentation is the recurring denunciation and discovery of the need for context and resolu- tion in index entries." 221 and Artandi and Hines who suggest that coordinate indexing "is moving toward the use of conventionally formed subject headings on the simplest possible level which can serve for the identification of a distinct subject and then applying bibliographic coordination of these headings." 222 3.6.3. Cooperative and Centralized Cataloging The history of cooperative cataloging in this country had its beginning when "as early as 1851 Charles Coffin Jewett proposed an organization of libraries in the country with the Smithsonian In- stitution as a center which would include coopera- tive cataloging among a group of cooperative en- terprises. . . . The Smithsonian officials declined to develop the program . . . The organization of the American Library Association in 1876, later developments in the compilation of Cutter's Rules, and the arrival of the 3x5 inch card had much to do with the picking up of the proposals." 223 Centralized cataloging, a form of cooperation, is defined by E. Thompson in her glossary, (1943 [569]) as "1. The preparation in one library or a central agency of catalogs for all the libraries of a system. 2. The preparation of catalog cards by one library or other agency which distributes them to libraries." She also defines cooperative cata- loging as "The production of catalog entries through the joint action of several libraries, in order to avoid duplication of effort. Particularly, the plan by which cooperating libraries prepare copy for catalog cards to be printed by the Li- brary of Congress." 224 In its position as the largest library in the United States, the LC has been influential in affect- ing the manner with which many libraries here and abroad perform their tasks, this influence tending to promote cooperative activities among libraries. One of the first evidences of such in- fluence was the issuance of printed catalog cards by LC for sale to the general public about 1901. The format of these cards including the suggested classifications according to both the Dewey Deci- mal Classification and the LC schemes helps li- brarians to keep records of what they have and where they have them. As Tauber says, "In the United States, the Library of Congress has been a centralized agency for cataloging, in addition to serving as a base for cooperative cataloging .... 2M Anprell, 1954 [331. P- 195-196. 221 Herner, 1962 [255], p. 6. 222 Artandi and Hines, 1963 [351. p. 76. 223 Tauber, M., I960 [563], p. 191. 224 Thompson, E., 1943 [569}, p. 26. The H. W. Wilson Company, which distributes cards to libraries on a more restricted basis than the Library of Congress, is also a centralized cata- loging agency." He continues, "Centralized cata- loging was started by the Library of Congress in 1902. A cooperative arrangement for acquiring card copy from other government departmental libraries was also put into effect. In 1902 there were 212 subscribers to the printed cards, return- ing a total of $3,785.19 to the United States Treas- ury." 225 The importance to librarians of ac- quiring these printed catalog cards is shown by the insertion of the LC card number in many books now published. In April 1930, LC began to add Dewey numbers to its printed catalog cards. Recently an interesting but unsuccessful experi- ment, called "cataloging-in-source," was tried by the Library. It involved reproduction of the ac- tual LC card on the verso of the title page of new publications. Publishers furnished LC with page proofs and data sheets for their publications, LC was to catalog the titles, and the publishers were to print facsimiles of the LC cards in the published book. Libraries were then to transfer the catalog entries from the books to cards for filing, by means of special camera equipment. Although the ex- periment was subsequently abandoned, Evans sug- gests that the difficulties which were encountered could be overcome and that an effort should be made to extend the practice to a wider range of publications both here and abroad. Such a sys- tem could sharply reduce the cost of cataloging in the world's libraries. 226 Types of publication which raise problems for catalogers in many other libraries as well as LC include technical reports and serial technical pub- lications. "The problem of cataloging and classi- fying the mounting volume of [technical] reports (some carrying a security classification) originat- ing in countless sources and varied forms outside the conventional channels of publication calls for extensive cooperation by federal and non-federal libraries .... A centralized mechanism to deal with these reports should certainly be estab- lished. . . . The total task is too huge to be ac- complished by any single library, but a great deal could be achieved by a coordinated system of indexes to which many libraries would con- tribute," 227 Evans continues, "The Library of Congress has long been active in the preparation of 'analytics' — card entries for items in periodicals and other pub- lications. This service might well be expanded on a government- wide basis as part of a major pro- gram which is needed to index and abstract the rising flood of serial technical publications." 228 If cooperative cataloging could not be extensive, an abbreviated record short of a full catalog entry would help to relieve the pressure of accumula- 225 Tauber, 1960 [563], p. 174-175, 191. 228 Evans, 1963 [188], p. 21. 227 Tauber, 1960 [563], p. 49. 228 Ibid, p. 50. 60 tions of unsorted and unidentified material which plague librarians. Even maintenance of an author card file and shelving by author would help. But centralized cataloging such as is carried on by the Veterans Administration would seem to offer much promise. The VA, which has been compared to a city library with 450 branches in 50 states, operates a coordinated acquisition and cataloging service centrally and supplies new books and their catalog cards ready for filing. The operation has brought cataloging costs down to 35 cents per title, as com- pared to an estimate of gross personnel costs of $3 to $4 per volume. 229 Another figure for catalog- ing a nonfiction book shows variations from 35 cents to $6 per title. 230 Veterans Administration officials indicate their willingness to consider requests from other Gov- ernment libraries to use this service on a cost basis. In regard to this centralized operation combining purchase and analysis, Evans suggests" . . . the fuller use of acquisition and order records as the first step in cataloging; the present failure to ex- ploit this information is most wasteful." 231 Another Government agency taking advantage of coordination in cataloging is the Clearinghouse for Federal Scientific and Technical Information which, since the inception of its cooperative pro- gram with the Defense Documentation Center, has made some changes in cataloging practices to avoid duplication and achieve economy. The Clearing- house uses DDC descriptors in its subject index- ing and produces paper tapes for input to the DDC computer as a by-product of its own an- nouncement function and catalog card prepara- tion, modifying an earlier arrangement between the predecessor agencies, Office of Technical Serv- ices and the Armed Services Technical Informa- tion Agency. Two cooperative ventures in cataloging have been undertaken by the National Federation of Science Abstracting and Indexing Services: (1) installation of a bank and share arrangement for in-shop files, glossaries, instructions, etc. for use of other member services (e.g., CA microfilm of 60,000 cross references for its subject index, DDC frequency analysis of descriptors) and (2) set up of a warning system for changes in format and coverage detail in existing systems of members, to encourage overlap and continuity between mem- bers in indexing. 3.6.4. Cataloging Rules and Guidelines We shall review, roughly in chronological order, the concerted activities in the United States look- ing toward rules and directions for the physical and content identification of books and other pub- lications. This is perhaps the area where the greatest amount of cooperative activity is needed and has been accomplished. Charles A. Cutter strongly influenced the sub- sequent cataloging activities of librarians by his 1876 Rules for a Printed Dictionary Catalog ([153]) as claimed by Landau, this is probably the first complete code of rules for a dictionary catalog and is still the basic standard code. 232 Among the first of the achievements of the ALA that are contributions to cooperation and com- patibility was the issuance of a guide in three suc- cessive editions : Condensed Rules for an Author and Title Catalog, 1883; Cataloguing Rules: Au- thor and Title Entries, 1908, issued by the ALA and the Library Association (Great Britain), the first international catalog code; and ALA Cata- loguing Rules for Author and Title Entries, 2d ed., 1949, [19]. During the early 1950's there was considerable thought given, around the world, to the adequacy of cataloging codes, and at the 21st annual con- ference of the Graduate Library School of the University of Chicago in June 1956 a series of papers devoted to this subject was delivered. 233 Angell took the view that the catalog should be a means of access not only to printed materials but to all the preservable manifestations of human communication, from manuscripts to sound record- ing to sculpture. He questioned whether the next coding system could be no longer traditionally book-oriented but equally suitable for all forms of records. Osborn, representing the group ap- pointed by the International Federation of Li- brary Associations to promote coordination in cataloging, spoke principally on those develop- ments which affect the two great cataloging tra- ditions, stating that the real trouble spot was the corporate entry. Chaplin spoke on a universal cataloging code and related the events leading to the decision by IFLA, at its Zagreb conference in 1954, to set up a working group on the interna- tional coordination of cataloging principles. Effort is being expended by representatives of the ALA and LC to prepare a new code of cata- loging rules as is shown by the tentative publica- tion in 1960 by the ALA of Code of Cataloging Rules: Author and Title Entry [349]. Subtitled "An Unfinished Draft for a New Edition of Cata- loging Rules," the publication was prepared by Seymour Lubetzky of LC for the Catalog Code Revision Committee of the Cataloging and Classi- fication Section of the ALA Resources and Tech- nical Services Division. The draft is accompanied by an explanatory comment by Paul Dunkin who claims that Lubetzky "rescues" the library pro- fession with his code and that Lubetzky begins by restating objectives which C. A. Cutter had stated in 1876. The objectives Lubetzky sets forth are (1) to facilitate the location of a particular publication, i.e., of a particular edition of a work, which may ~» Evans. 1903 riSSli, p. 4. 230 Oellrich. 1903 [435], P- 67. 231 Evans, 1963 [188], p. 25. 232 Landau, 1958 [3271. p. 65. 233 See The Library Journal, October 1956. 208-371- 61 be in a library ? and (2) to relate and display to- gether the editions which a library has of a given work and the works which it has for a given au- thor. The draft of the Code contains four parts : works of personal authorship ; works of corporate authorship; government publications; and works of unknown, complex, or changing authorship. There have been important changes suggested in the rules for the Code. As of 1961 most li- brarians were ready to recommend use of the new rules primarily for new authors with only limited effort to be made in changing established headings, but in general there is still no agreement on rules for descriptive cataloging in a dictionary cata- log : ". . . the library profession, since the appear- ance of the ALA revision, has not agreed on its cataloging rules." 284 A booklet containing instructions for the cata- loging of technical reports and related materials has been made available by an agency of the U.S. Government. This is the DDC's AST I A Guide- lines for Cataloging and Abstracting, which also contains instructions on matters other than de- scriptive cataloging. Joint meetings on descriptive cataloging were held in 1963 by representatives from DDC, NASA, OTS, and the National Bureau of Standards. One result of these meetings was the Standard for De- scriptive Cataloging of Technical Reports, needed for the effective retrieval of reports as reference material. Such use was hampered by the wide variety of labels and codes attached to them in great but seemingly meaningless array. Taube in 1950 discussed the cataloging of pub- lications of corporate authors, stating that the rules for corporate entries in the Science and Tech- nology Project of LC began with the specification that ". . . corporations or other collective bodies of men under whatever name are to be considered and treated as authors of their official reports, proceedings, and other publications for which they are collectively responsible. . . . Enter the publications of a corporate entity under its name. . . . The form of the name is to be determined by information available in the work being cata- loged and in authority lists from cataloging pre- vious works and from these two sources only. The form of the entry given on the title-page or else- where in the work being cataloged may be modi- fied in the entry only if a different form has al- ready been established and is used in the cata- log." 235 Such a rule relieved the cataloger from performing "research work" or searching to dis- cover other possible and perhaps more accurate forms for entry. A third rule concerned the sub- divisions of corporate bodies whose names are not suitable for entries: "If the title-page or other parts of the work being cataloged disclose that a division or part of a corporate body was respon- sible for the report, entry should be made under the division or part." 286 These rules are still advocated, as suggested re- cently: "Any new code developed by agencies working in the report literature should at least take these trends [cataloging from the information contained in the document itself, use of name most frequently used in publications rather than offi- cial names, rejection of distinctions between types of corporate bodies and of entry under geographic location, emphasis on entry under most specific entity, etc.] into account. They are the result of a great deal of work by experienced and progres- sive librarians engaged in a needed reform activ- ity, and they are undoubtedly the pattern for fu- ture international growth in this area. Any given report-processing agency might be either too small, with a custom-built system operating reasonably well, to consider any drastic changes ; or it might be too big, with too much invested in existing card catalogs, tapes, indexes, etc., to consider changing. But those not in these categories, agencies just starting their activities, or those who must arrange for consistency or compatibility among various agencies, could with profit study these develop- ments." 237 Descriptive cataloging rules concentrate on physical identification of publications; content identification also raises problems and needs prin- ciples for guidance. "Subject cataloging . . . has many ramifications. It is not suggested that a tight standard be attempted but rather a general guideline standard with the requirement that sub- ject cataloging be done in depth on each technical report. It is suggested that the professional li- brian of the activity originating the report set forth the subject cataloging entries in a descend- ing order of importance as they pertain to the report. This will allow a receiving library to scan the entries and select those used in their own li- brary subject catalog." 238 The International Conference on Cataloging Principles, sponsored by IFLA, was held in Paris in 1961. The preliminary official report of the Conference was published by the UNESCO Bul- letin for Libraries (1962 [286]). The official re- port was published in 1963 by the International Federation of Library Associations in London and was made available in the United States in 1964 by the ALA, Chicago. Atherton, an observer at- tending the Conference, reports that the delegates "discussed the fundamental principles underlying the choice and form of headings and entry-words in alphabetical author-and-title catalogs. . . . The discussions about corporate authors (what works should be entered under corporate authors; form of heading for corporate authors, and subordinate bodies) showed quite clearly that it was going to be possible to have greater agreement in library 234 Commerce by Taube, Conference Session II, In Markuson, 1964 [372], p. 58. 2 ^ Taube, 1950 [556], p. 15. 23 ° Ibid, p. 18. 2 " Brandhorst, 1964 [76], p. 42. 238 Langenbeck, 1962 [329 J, p. 298. 62 catalogues of the future than it could have been hoped possible." 239 Resolution IV called for the compilation by each country of lists of the names of the principal cor- porate bodies existing in each country, for the purpose of standardizing corporate author cataloging. 240 3.6.5. Use of Machines for Description and Identification There is a dearth of mechanical devices, multi- purpose forms, and other means of reducing the clerical work connected with cataloging. For in- stance, catalogers could record most of their deci- sions on a dictaphone or Stenowriter. The tape typewriter, of which Mooers spoke so eloquently in England, France and Germany in the Spring of 1960 ( [389] ) , is one tool that can be used for cata- log production. The tape typewriter plan offers means for ( 1 ) improving cataloging and eliminat- ing duplication of effort in cataloging among libraries and (2) increasing clerical efficiency in the cataloging operations now being done at each library. The tape typewriter is described as having a keyboard whose operation causes both the charac- ters to be imprinted on a sheet of paper placed in the machine and a perforated paper-tape record to be made of everything typed. The typewriter must also have a "reader" to read the perforated tape and operate the typewriter in accordance with the perforations on the tape. Mooers states that "in a fully capable tape type- writer the letters of the alphabet, numerals, punc- tuation, 'capitals,' 'carriage return and line ad- vance,' 'tabulate,' and 'backspace' are all recorded by patterns of punches in the tape." Once the paper tapes have been read into the memory of an electronic computer, "the machine can rearrange the sequence of entries making up each card, put- ting (for example) the author's name first. Then inside its electronic and magnetic memory, it can also perform an alphabetization by author's name for a large collection of cards. When these alpha- betization operations are completed, the machine can then, perforate a new tape which will control a typewriter to type up cards in alphabetical sequ- ence according to the authors' names !" 241 "The basic principle which is being stressed by Mr. Mooers in his tape-typewriter plan is that any mechanical storage which is accomplished should be done in standardized machine-readable form so that the machine-readable text can be used by many persons and installations throughout the world." 242 "This plan offers several advantages to the re- trieval center and cooperating institutions in addi- tion to the speed of manuscript transmission via paper tape. First, this machine-readable record of the bibliography or portions thereof, will be ™» Atherton, 1961 [40], p. 583. »« See Brandhorst. 1064 [761:, p. 38. J « Mooers, 1960 [393], p. 277-281. J «Gull, 1962 [228], p. 58. made available to cooperating documentation units and the possibility of reciprocal exchange exists. Also the preparation of edited machine-readable bibliographic entries is done once, never having to be repeated. "The first large scale implementation on an in- ternational basis of the 'tape-typewriter plan' for cooperation between documentation centers, has been initiated between Italy and the United States. This particular application of the plan entails the transmittal of entries in an international serial current bibliography from the . . . FAO . . . headquarters, Rome, to the Aquatic Sciences In- formation Retrieval Center (ASIRC) at the University of Rhode Island." 243 Within one to two years of Mooers' announce- ment, Bernstein (1962 [57]) of Euratom/CETIS, Brussels, Belgium, was writing in the German Nachrichten fur Dokumentation and in the UNESCO Bulletin for Libraries that generally a Flexowriter is economical under the following cir- cumstances: (1) where a number of different doc- uments must be processed; (2) where the docu- ments must be processed by a routine that is in- variable or only slightly variable; (3) where the routine can be divided into stages for successive processing; and (4) where information from pre- vious stages is repeated at these stages, other infor- mation added, and part of the resultant total body of information transmitted to the next processing stage. Lipetz describes techniques and equipment for simultaneous production of manual and machine- usable book cataloging records as a first step in introducing a machine record system for the book catalog at the Research Library of the Air Force Cambridge Research Laboratories, L. G. Hanscom Field, Bedford, Mass. Punched paper tape was created in the course of continued production of 3x5 inch index cards as similar as possible to those in the existing catalog. The input typing format, designed for a special-purpose computing machine, the Crossfiler, produced sets of catalog cards at less cost than the utilization of a general-purpose computer. Sets of catalog cards were obtained with identical text in the body of a card but differ- ent tracings (entries at the top of a card for title, abridged title, subject, etc.), derived automatically from the text contained in the primary punched tape. 244 An example of an early investigation into mech- anized documentation is the activity preceding the arrival of MEDLARS at the National Li- brary of Medicine, the "Welch Medical Library Indexing Project," a research program conducted by the Welch Medical Library of Johns Hopkins University and supported by the Armed Forces Medical Library through a contract between it and the University. Larkey reported (1953 [332] ) that the Project began with the interest of 2 « Scientific Information Notes 5, 3, 1 (1963). 214 Lipetz, 1963 [343], p. 119, 121. 63 the Army Medical Library in the program of the Welch Library (following World War II) for training medical librarians in the necessary re- search aspects. The scope of the project included (1) studying the problems of indexing medical literature; (2) exploring the theory and practice of subject heading (nomenclature) and classifica- tion (coding) as they concern medical literature; and (3) exploring methods, particularly machine methods, applicable to medical bibliography, oper- ating any pilot projects needed, and reporting on suitability of machine methods in the operation of the Army Medical Library. Larkey reported at the 1956 [331] "Conference on the Practical Utilization of Recorded Knowl- edge — Present and Future" on the prospects for cooperative information processing in medicine. He suggested that until that time there had been little effective cooperative effort. In the 1920's the American Medical Association, publisher of Quar- terly Cumulative Index, and the Surgeon General's Library (later Armed Forces Medical Library, and still later the National Library of Medicine) , which sponsors the Index-Medicus, cooperated to compile a merger, the Quarterly Cumulative Index Medi- cus. In 1956 the AMA became solely responsible for the Quarterly Cumulative Index Medicus, and the Armed Forces Medical Library revived the Current List of Medical Literature, begun in 1941, destined to become the new Index Medicus. Larkey also saw a need for "full and frank dis- cussion as to how and why the various information aids are produced ; why they select material from certain documentary sources (periodicals and other sources of information) ; then, in turn, why these documentary sources exist, for what purpose, and their relation to the local research effort." 245 He identified an additional area for cooperation in the preparation of the subject indexes for the Current List or Chemical Abstracts by machine methods, with the modification of the "coding as used for the index to coding for machine searching." Larkey related that the Welch Library Indexing Project had found real possibilities on the basis of experi- mental runs. "Here we made use of the subject entries and subdivisions from the actual subject in- dexing slips for an issue of Current List. It might be advisable to code for information, as from the 'modifications,' as in Chemical Abstracts and Cur- rent List? 246 The MEDLARS project, current work on mech- anized indexing at the National Library of Medi- cine, has already been described briefly. The sub- ject headings published in Medical Subject Head- ings "are the only tags that may be used as head- ings in Index Medicus. ... In effect, there is one level of indexing for Index Medicus and two levels of indexing for both demand and recurring bibli- ographies. ... It is anticipated that each cita- tion will have an average of 10 tags assigned, but 245 Larkey, 1956 [331], p. 304-305. 246 Ibid, p. 306. only an average of three of these tags will be checked by the indexer as those tags under which the citation will appear in Index Medicus? 247 Other tags prepared by the indexers and never appearing in Index Medicus will include provi- sional headings for new descriptors, age group tags, Public Health Service support group tags, and geographical tags. In the case of Nuclear Science Abstracts, Lay and Lebow describe a procedure which began in 1959 to provide four indexes for subject, personal and corporate authors, and report number (with availability information) for each semimonthly issue of the AEC's NSA. "The new AEC tech- niques have eliminated the time consuming and costly manual operations by using speedy and eco- nomical machine operations. The new system uses equipment that is readily available and is suc- cessful through the effective adaptation and com- bination of the products of several manufacturers. In its simplest terms the index production system provides for : (1) typing of index entry in a pre- scribed area of an IBM accounting machine punch card; (2) machine coding the punch card so that the thousands of index entries can be machine sorted and arranged; (3) machine controlled photographing of the cards at a rate of 230 cards (lines) per minute; and (4) arrangement and makeup of the resultant negatives into plate nega- tives for photo offset printing." 248 Day and Lebow suggested that there would be no need for AEC catalog cards thereafter : "Each user of Nuclear Science Abstracts, whether at the laboratory bench or in the library, will have in effect a complete catalog to nuclear science infor- mation." 249 Sherrod, however, sees that ". . . the elimination of one requirement introduces a new requirement. Specialized bibliographies with ab- stracts no longer can be prepared easily by manual shingling of catalog-abstracts cards. Thus, there is a new requirement for storing abstracts in such a way that when special bibliographies are provided on request it will be a simple matter to include an abstract with each entry without going through the laborious process of retyping." 250 The AEC Division of Technical Information Extension has experimented with tape typewriters for producing technical abstracts and also with "semiautomatic indexing" from the information contained in an abstract and from a single effort in the production of that abstract for hard-copy purposes. The experiments on indexing are re- ported and demonstrated in RDA-3 and -4, the Research and Development Abstracts of the US- AEC for July-September and October-December 1962. _ The reasoning of the AEC, as explained in these issues, was "that, if the title plus the abstract could be used as a source of subject headings, the depth and range of subject indexing would be con- 247 National Library of Medicine, 1963 [412]^ p. 9-10. 218 Day and Lebow, 1960 [161], p. 122. 249 Ibid, p. 127. "0 Sherrod, 1963 [529], p. 215, 64 siderably increased over that possible with KWIC indexing from titles only. . . . Words or phrases were underlined by the abstract editor to provide suitable subject points. These terms were then coded in tape by the Justowriter operator at the time the abstract was prepared for the journal." The method was "devised to code or 'lock' into the tapes the various components, terms, phrases, etc., of the abstract that would produce the follow- ing indexes: (1) personal author, (2) corporate author (source), (3) subject, (4) report number, and (5) journal citation." In a paper describing the project, Sherrod concludes that the economic feasibility and other technical advantages of this method over former conventional procedures was proved : "What remains to be determined are the retrieval efficiency and/or the degree of user ac- ceptance of this new subject index." 251 A most significant development in the area of achieving cooperation, compatibility, and con- vertibility by virtue of author-editor-machine col- laboration in the indexing of journal articles and conference proceedings is the FASEB experi- mentation with machine-aided compilation and normalization of indexes. In a sense, this is the first example of a computerized authority list; in another, it is a practical application, at least in a prescriptive manner, of a mechanized thesaurus. Schultz and Shepherd have described (1960 [512] ) the use of the UNIVAC computer in sched- uling the meeting of the Federation of American Societies of Experimental Biology (FASEB) in April 1960 and preparing an index for the meet- ing. During the summer of 1959 the 2,383 ab- stracts of the 1959 meeting were indexed; a com- puter was used to determine those index terms with enough specificity to classify the abstracts into groups of 10 or 12 each. The computer findings were then modified and rearranged to produce a list of subject categories designed to make explicit the areas of current researcli in experimental biology of interest to FASEB. The UNIVAC computer and other equipment were then used (1) to publish the 2,526 abstracts submitted or spon- sored by members in 1960; (2) to arrange the ab- stracts into subject groups of 10 or 12 each for purposes of programming the meeting; (3) to schedule the sessions to avoid conflicts whenever possible; and (4) to prepare a subject index to the abstracts. In 1963 Schultz ([509]) used a magnetic-tape thesaurus and a computer program to edit author- produced indexing terms and phrases for the FASEB Proceedings of that year. The FASEB thesaurus was constructed after considerable per- usal of other authority lists in the biomedical field. For input to the computer routine, there were the key-punched title to each paper to be indexed and the indicia supplied by the author, the words thereof being marked "with the number of the document from which they originate as well as the 251 Sherrod, 1903 [529], p. 215. number of the sentence or phrase and the position within the sentence or phrase." After the mean- ingless words were extracted from these sources, the "remaining words are resorted into their ori- ginal order within each document description and all sentences and phrases are sorted into alpha- betical sequence. Processing against the thesaurus is begun on a word-by-word basis for each of the sentences and phrases." This computer program uses the thesaurus to perform four basic functions: (1) it accepts cer- tain words without modification; (2) it recognizes certain words as being acceptable if modified, and then it modifies them; (3) it recognizes certain words as acceptable and adds other indexing terms on the basis of their presence, (thereby perform- ing a "syndetic" function, i.e., automatically post- ing "see also" terms) ; and (4) it deletes certain nonindexable words not on the original delete list because of the possibility of their being contained in an acceptable indexing phrase. Three other citations are of interest in the matter of application of computer techniques. Welt in speaking about indexes and index mechanization in biomedicine says, "It is now possible to provide general, shallow indexes to the entire field as well as specialized, deep indexes tailored to the needs of specialized research workers or practitioners." 252 Taube is convinced that "the design and display of terms in an authority list or 'thesaurus' to be used for both machine searching and as a guide to a published index remain to be determined. Experience may indicate that the two purposes cannot be combined and different code books of 'thesauri' may be necessary for the mechanized retrieval portion and the published portion of an information center's apparatus. It may also turn out to be the case that different types of guides and instructions are required for the indexer and searcher." 253 Weber, writing in 1957, decided in considering the quagmire of scientific literature that "the con- clusion to be drawn is that the use of machines for storage and retrieval of information is likely to be practicable only through a man-machine part- nership, and is not going to be commonly feasible for many years to come." 254 He says specifically concerning indexes that "it is logical to expect that a great increase in extremely brief subject entries, arranged in chronological order, will characterize the future subject indexes to scientific material — with the older material being indexed merely by an author file, and with subject cards thrown out after a period of time." 255 Machine use in the generation of assignment indexes implies requirements for compatibility and convertibility for mutual agreement as to contents and scope of commonly used technical thesauri or mechanized authority lists ; mutually agreed upon 252 Welt, 19(63 [641], p. 173. 253 Taube, 1963 [558], p. 179. 254 Weber, 1957 [631], p. 100. 255 Ibid, p. 118. 65 listings of equivalent index term vocabularies for interconversion, one system to another ; and specific identification of format, field designators, and coding practice if all machine-usable records are to be interchanged. "For automatic indexing, statistical methods based on frequency of words in documents are being investigated. Similar procedures are also being tested in the production of 'automatic ab- stracts.' The effectiveness of such procedures in selecting the significant portions of scientific papers has yet to be demonstrated. . . . Further research is needed to determine whether automatic indexing and abstracting procedures and nor- malized languages will be of practical use in storage and retrieval systems." 256 3.7. Systematization and Terminology Control Enforcement of both lexical and syntactic regu- larities in indexing vocabularies and documentary languages requires either compatibility or con- vertibility if processed materials are to be ex- changed between two or more systems. In addi- tion, compatibility or convertibility within a single system is required if the searcher or user is to be freed from rigid adherence to the prior indexing rules and language. Two rather comprehensive examples of such doc- umentary languages are the Western Reserve semantic code and telegraphic abstract system and SYNTOL (/Sfynfoigmatic Organization of Zan- guage). In both these cases, some consideration has already been given to mechanized procedures for automatic conversion from ordinary language to the system language. Convertibility between the language of search requests and the selection indicia within these systems is achieved by consist- ent application of the translation rules both to the language of the requests and to that of the items stored. That is, a translation is made from ordi- nary words in natural language text to regularized words or terms in a documentary language. Standardized, or mutually agreed upon, vocab- ularies at a variety of levels are required: (1) at the level of code representation of characters in machine language: number of characters accom- modated by code, which characters are accommo- dated by code; (2) at the level of characters ac- ceptable as input: size of allowable vocabulary, what characters ; provision for other characters as required for multiple-coding, shift keys, etc.; (3) at the level of characters acceptable as output ; (4) at the level of indexing vocabulary or documen- tary language; (5) at the level of aids, e.g. word- lists, thesauri, dictionaries, stop lists, and the like, and (6) at the level of internal processing of character strings: fixed versus variable word length, fixed versus variable field length, fixed ver- sus variable record length, and fixed or variable file length. 2M U.S. Senate, 1960 [592], p. 105. Next, and more crucial, is the level of the docu- mentary language itself "There is no doubt that, if rational methods of automatic information re- trieval are to be established, a study of the lin- guistic structure of documents from which the information is to be drawn cannot be neglected — regardless of the particular method used. Fur- thermore, one of the problems which documenta- tion must certainly solve in the future (and per- haps its most important problem) would seem to be that of devising a rational 'documentary lan- guage' standardized on an international scale." 257 Even though such study of linguistic structure seems necessary, "there have been relatively few studies of principles of organizing or classifying knowledge systematically and few attempts to de- vise better systematic organization irrespective of the techniques (manual or mechanized) subse- quently used in processing information." ^ There have been developed, however, "... a cer- tain number of rather general practical procedures, resulting from common requirements and condi- tions which are imposed upon them by character- istics peculiar to the new 'machines' (this term is here used in its broadest sense) for the retrieval of documents and the information contained therein. Such were, for example: the breakdown of com- plex subjects into simple factors (or, at least, less complex) ; the transition from highly hierarchical classification systems to less hierarchical systems ; the realization of the necessity of expressing not only the elementary terms used in the analysis of documents, but also the relations between them." 259 Early experiments with such developments are found in the work of Perry, Berry, and Kent. The capability for using hand-sorted margin- punched cards for small collections of information had been reported in 1945 by Cox, Bailey, and Casey ([145]). In 1946 the Board of Directors of the American Chemical Society set up a Com- mittee on Punched Cards with James W. Perry, then a library fellow at the Massachusetts Institute of Technology, as head. In 1947, a modest sum was obtained from industry to continue the invest- igation. Following the expiration of the ACS grants, the support of the activity was taken over by the MIT Center for Scientific Aids to Learning, with the aid of a grant from the Carnegie Cor- poration, Perry continued as chairman of the re- named ACS Committee on Scientific Aids to Lit- erature Searching and at this time he and G. Mal- colm Dyson approached Thomas J. Watson, Sr., president of IBM, suggesting the development of machines suitable for searching literature. H. P. Luhn, of IBM's engineering laboratory, then proposed sending the standard IBM punched card through a machine endwise, devising a punch code by which any information could be punched in any of the 80 vertical columns, and using photo- electric scanning and the principle of comple- 257 de Groller, 1962 [170], p. 144. 258 U.S. Senate, 1960 [592], p. 103-104. 269 de Grolier, 1962 [170 ]i, p. 10. mentary pattern matching: To look for a given piece of information, "the complement of its code — a 7-hole pattern — is punched in a question card. The question card is mounted in the scanning ma- chine so that the cards to be examined pass between it and the photoelectric cell. When a match is made, no light passes, and the punched card is kicked out." 260 Perry and his associates then began work to de- velop a machine language for literature searching (1956 [455]). They centered their attention on the semantic aspects of machine searching, con- cluding that the analysis of index terminology could be incorporated into a code dictionary since the semantic significance of most words for scien- tific and technical indexing remains reasonably invariant in different contexts. Perry and his as- sociates arrived at the concept of semantic factor- ing where words are examined for their semantic components. (A colorimeter is a machine for an- alysis with color, so its code designation must refer to all these factors.) Perry also developed "role indicators" whereby there is affixed to the codes for various terms special symbols for designating the function or role. In this way distinction be- tween initial reactants, final products, and condi- tioning agents is possible by preceding their code designations by S, F, and K, respectively. These investigators proposed four steps for processing information for search by automatic equipment : ( 1 ) underlining keywords and phrases or supplying marginal notations to designate im- portant aspects of a subject; (2) editing the under- lined or annotated words or phrases to produce an "abstract" of unusual literary style which, how- ever, is intelligible before encoding; (3) encoding the edited abstract with the aid of the code dic- tionary; and (4) recording the encoded abstract by means of a card punch. The fourth step may be eliminated when encoding will be accomplished with an automatic, key-punching dictionary. The general approaches to standardized lan- guages or vocabularies for information systems are concerned with thesaurus development, with classification schemes, and with interchange of in- dexing languages. We shall examine these in turn. 3.7.1. Thesauri The potentialities of thesaurus-aided systems for indexing, abstracting, storing, selecting, and re- trieving pose both new requirements for coopera- tion, compatibility, and convertibility, and new opportunities in each of these areas. For the most part, the early proposals for a thesaurus-type ap- proach to mechanized documentation envisioned manual compilation and preparation and manual scanning for either indexing or searching pur- poses. It was quite generally assumed that major cooperative efforts between representative users and subject matter specialists would be required to 200 "Now Tools for the Resurrection of Knowledge," 1954 [423]i, p. 867. organize such tools. More recently, it has been suggested that machines can be used to assist in construction, and updating or reorganization, and to provide an "early-warning" system with respect to changing terminology or to changes of subject interest in a collection. Examples are "words not found in the diction- ary would be noted by the machine during the encoding process and spelled out for subsequent manual editing. This would involve the assign- ment of a notional family or families to the new word and its addition to the index. As a matter of course, each new document would also be sub- mitted to a search operation to discover duplica- tions, variations, and contradictions with respect to previously stored documents. This search would also serve to obtain a measure of the dis- criminatory characteristics of the system as the collection grows." 261 "If it is assumed that the documents in a typical technical library reflect the purpose of the library reasonably well, then it is logical to allow these documents to generate both the indexing diction- ary and the thesaurus. One way of doing this is to prepare a word-association matrix of the type developed by the Integrated Engineering Control Study of DuPont. This is essentially a list of asso- ciated terms statistically generated from the docu- ments in the DuPont store. It is primarily a thesaurus used as a word-reminder list, but tends also to generate connotative definition of words in the sense that two words almost always found in the same document reflect on the meaning of each other." 262 "Use of a thesaurus in editing the input data for the index is an optional feature of the generalized indexing technique . . . Once the thesaurus has been constituted, it is convenient to use it as a means of proofreading the input descriptors. Thus, every descriptor not appearing in the the- saurus ( and therefore not recognized by the index- ing system) is put out for human inspection. Discrepancies . . . may be either mispellings . . . or new descriptors not yet recognizee!." 263 Even in the early proposals for thesaurus use in documentation a number of aspects likely to involve machine considerations of compatibility and convertibility were raised. Some illustrative examples are as follows: ". . . It is reasonable to think that the further we can go in routinizing and mechanizing the techniques of translating ordinary language into a regularized language and of coding for machine manipulation, the more will we be likely to achieve economically feasible ma- chine searching on a large scale . . . [Some investi- gators] have come up with the thought that the best answer . . . may be the application of a mechanized thesaurus based on networks of related meanings. an Luhn, 1057 [3G0], p. 316. ** Lockheed Aircraft Corp., 1959 [348], p. 34-35. 2M Schultz and Schwartz, 1963 [511], p. *» Brownson, 1957 [83], p. 99-100. 422. 67 "Indexers and searchers have in the past got along without such a tool — or, rather ? each has used the network of word associations m his own mind, aided perhaps by a dictionary, as a personal thesaurus. It is in contemplating the possible mechanization of indexing that the need for this tool has been recognized. Passage from text- words to key-words can be achieved by machine only if such word associations have been previously stored in its memory." 265 The question of where and when the use of a thesaurus-type device to provide greater vocab- ulary compatibility or convertibility was first proposed, and by whom, is an elusive one. Vickery writing in I960 [614] on "thesaurus" as a new word in documentation, discusses prior usage by Brownson (1957 [84], Luhn 1957 [360], 1959 [352]), Bernier and Heumann (1957 [56]), and Taube (1955 [561] ) , the latter for the idea of both indexers and searchers referring to word-associa- tion lists. Vickery further discusses usage of the word by Americans, ignoring the emphasis in the Cambridge Language Kesearch Unit on thesaurus- like approaches to both mechanized translation and information retrieval from at least as early as 1956 onward. 265a T. M. Williams in a private communication to H. L. Brownson dated August 2, 1961, cited her own use of the word in the Avion report of No- vember 1956 (Williams, 1956 [648]) and of less formal usage both in oral presentations and in cor- respondence with Bernier during 1953-54. Wil- liams also notes that Bernier may well have im- plied the functional concept, although without the use of the specific term, in unpublished papers given in 1953 and 1948. Luhn is credited by de Grolier with having ad- vocated, at least as early as 1951, a thesaurus method : ". . . The process of 'broadening the con- cept' which he described was linked, in an interest- ing manner, to research strategy, and he proposed the compilation of a dictionary of notions where each of the specific terms not retained for index- ing would appear under as many headings as nec- essary to represent it by the different 'key terms' where it would thus be indexed. Such a diction- ary is very similar, in fact, to the Thesaurus of English Words and Phrases by Peter Mark Ro- get . . . and since that time it has been the custom to refer to the thesaurus method." 266 In November 1951, Mooers issued his Zator Technical Bulletin Number 65 (1951 [394]), not- ing in the preface that it was substantially the same as a paper informally circulated to members of the Punched Card Committee of the American Chemical Society in February 1947. In it, he states : "To surmount this problem of alternative expression, there must be a word book or encyclo- pedic source of volcabulary having features com- mon to a thesaurus, a dictionary, and an encyclo- pedia." 2G7 Finally, in addition to these "anteriorities" of discussions of proposed thesaurus-type functions without the use of the term and of linkages of the term to proposed functions, a reference appeared in 1954, in Appendix III of the Bush Committee Report: "Then there might be a large index and cross-reference file, adaptable to frequent addition, revision and reorganization. This file should be capable of rapid data processing and access. It might include a thesaurus so that the processing machine itself can translate search inquiries from a variety of sources into a standardized machine language, and so that changing terminology may be readily incorporated into the search and re- trieval procedure." 268 The term "thesaurus" has also been applied to the somewhat narrower concept of an authority list incorporating syndetic devices including "ge- neric to" or "specific to" cross-references. Such tools are provided by AIChE, by DDC, and by E JC, in their respective thesauri. "Probably the most outstanding [of the solu- tions developed for control of synonyms] is the thesaurus approach, which has been very success- fully used by Dr. Fred Whaley at Linde Air Prod- ucts Laboratories . . . and by the Engineering Information Center in . . . the DuPont company. Thesauri may be created in several ways, one of which is the Committee approach to consideration of vocabulary, a typical product of which is iden- tification and definition of synonyms and near- synonyms . . . The thesaurus approach can solve the problem of generics as well as the problems of viewpoint and semantics. The indexer can refer to his thesaurus for the terms he has selected and from the appropriate lists under each term, select those associated terms which refer to generically higher concepts appropriate for indexing the in- formation more broadly." 269 Wall wrote about "information retrieval the- sauri" under the sponsorship of the Engineers Joint Council in 1962, saying that such referents as a thesaurus for information retrieval and Ro- get's international thesaurus serve their users "for similar immediate purposes and in approximately similar fashions. Both users wish to identify terms 'related' in some manner to a term or terms they have already thought of. Both users look up alphabetically the terms already in mind. . . . Each user observes that some of the related terms found are of interest to him at the moment and some are not. Each user may use some related terms which are of interest to him as 're-entry' 163 Vickery, 1960 [619], p. 183. 205a Masterman, 1956 [374] and Parker-Rhodes, 1956 [446]. 200 de Grolier, 1962 [170], p. 109, 187-188 cites Luhn (1953 [358]) as having first appeared as Appendix I (dated 10 Sep 1951) to an unpublished IBM report (Luhn, 1952 [356]). a» Mooers, 1951 [394], p. 14 208 U.S. Department of Commerce, 1954 [596], p. 63. It should be reported, however, that Stevens when compiling material for this Appendix had access to both the Mooers 1951 Bulletin [394] and to the Luhn 1952 report [356] the latter having been included In the list of selected references, p. 76. »» Costello, 1961 [138], p. 22-23. 68 points in his thesaurus to find, other terms of in- terest." 27 ° He wrote again about an information retrieval thesaurus, that "it must certainly list the terms of the system vocabulary. It must exhibit relationships among these terms . . . relation- ships such as synonymy, hierarchy, and relation- ships which may indicate synonymy or hierarchy from some points-of-view but not generally. It should define the vocabulary terms to the extent required." 271 Herner, writing about the possibilities for con- vergence of word and concept indexing through the intermediary of the thesaurus, expresses the hope "that designers and users of indexing vocab- ularies will recognize this — if they haven't al- ready — as a very basic purpose for thesauri, subject authorities, indexing dictionaries, or whatever name we give them." He gives examples where "thesauri and similar indexing authorities have been used as means of converting words or symbols and groups of words or symbols into concepts." 272 This, then, is the general setting for the current usage of the term in this country and by some of the agencies of the U.S. Government. Taube suggests that ASTIA and "several other organizations and individuals have continued to insist that in 'thesauri' there are provided struc- tures of connections between words which are necessary to ensure satisfactory operation of mech- anized systems of coordinate indexing." He then asks, "Is a thesaurus or any authority list an independent semantic standard for an indexing system set up by a process of lexicography, or is it a description of a particular indexing system as developed from a concatenation of subject com- petence, the literature being indexed, and the re- quirements for efficient machine search?" 273 Taube claims "that a cross reference structure of any complexity can be derived from, and need not be prescribed for, any given indexing system. Further, if ... a descriptive structure contains cross references which reflect the extensive rela- tion of classes of items rather than prescriptive semantic relations of terms, there is no real problem of compatibility between descriptive systems." 274 He suggests further that for revising its first thesaurus ASTIA should have printed out the terms and postings recorded during the indexing operation following the appearance of the edition. From this document there could "be derived all necessary 'see' references and references to indicate the inclusion of one class in another. For studying the relations which are basic to 'see also' references, it is necessary to have a printout of the tracings. ... If any two terms have identical postings, a 'see also' reference from one to the other is re- dundant and wasteful. In such a case, one set of 2TO Wall, 10fi2 [623], p. 7. "'Wall, 1963 [G25], p. 38. 272 Herner, 1903 [250}, p. 1S.3. 273 Taube, 1903 [558], p. 177. 2 " Ibid, p. 178. postings should be eliminated and a 'see' reference substituted for the 'see also' reference." 275 A final interesting comment is that of Cleverdon and Mills : "Work on thesauri and classifications, where it has been practical in nature, appears to consist of compiling lists of terms which go out of favor as quickly as any list of subject headings in the past . . . ." 276 Turning now to established thesauri, we note first the ASTIA thesauri. Heald states that the purpose of Project MARS (MAchine Retrieval System) was to prepare a thesaurus of scientific descriptors from the 70,000 ASTIA subject head- ings and to assign descriptors to all AD-numbered reports in the ASTIA collection. 277 The subject headings in the 1959 volume were divorced from their subdivisions, about 150 subdivisions were eliminated, and some subdivisions were included "in other headings as synonymous or definitive terms." 278 The almost 7,000 descriptors were "di- vided into a number of generic groups for display purposes. Hence, the full scope of the subject coverage is divided into about 290 major categories." 279 Slamecka comments on the grouping of indexing terms: "Classification may be exhibited in a thesaurus in two ways ; through relationships . . . among terms and through grouping terms into categories. The latter device is very helpful if the indexing environment is very broad, as it is in the subject field interests of ASTIA." 280 The number of "descriptor groups" in the first ASTIA Thesaurus is actually 292, and we shall see that they have been used in experimentation for con- verting one indexing language into another. The first edition of the ASTIA Thesaurus was revised in late 1961 and 1962 under a contract between the Engineers Joint Council and ASTIA, the second edition appearing in December 1962. Of interest here are the experiences of a worker with the first edition. Hicks reported her experiences with govern- ment-sponsored research reports in areas of physi- cal chemistry to the Division of Chemical Litera- ture of the American Chemical Society in May 1962. One of her searches was on the Dissociation and recombination kinetics of hydrogen, the halogens, and the hydrogen halides. Machine searches at ASTIA headquarters yielded 82 perti- nent ASTIA reports. Manual search was made of the subject files in the ASTIA Los Angeles 275 Taube, 1903 [558], p. 179. Note tbat in the chemical engi- neering thesaurus (to be discussed below in the section on AIChE and E.TC thesauri), the terminology "RT" for related terms is used instead of "see also." Taube (Ibid. p. ISO) re- marks that "Even authority lists for manual catalogs have difficulty with 'see also' references unless they are restricted to references from general to more specific classes ; and what emerges from this analysis is the thesis that the notion of 're- lated term' as an explanation of 'see also' references is too vague to be significant either to the indexer or the searcher in a mechanized system." 27B Cleverdon, 1962 [124],. p. 1. 277 Heald, 1960 [243], p. 117. 278 Ibid. 270 Ibid, p. 119. 280 Slamecka, 1963 [534], p. 227. 208-371- 69 Regional Office for reports from 1953 to the end of 1959, and selected sections were examined page by page. By manual searching, 242 reports were found. Three reasons were ascribed for the dis- parity in the number of items found: "(1) inac- curacies and insufficient depth of indexing at ASTIA; (2) manual location of relevant reports which we had not anticipated finding and which we had not asked the machine to find ; (3) less than optimum prescription of retrieval terms on our part." 281 Hicks' second example was Vacuum, ultraviolet photolysis of organic compounds. She manually searched TAB for 1960-1961, finding 23 reports on vacuum ultraviolet photolysis of various com- pounds. She comments that it was apparent "that the easiest way to assuring nearly complete re- trieval of these reports would be to prescribe one descriptor only, Photolysis, requiring manual screening of at least 112 cards or examination of at least four times as many cards as those found pertinent. Because of underassignment of terms, complete retrieval of these 23 reports would re- quire programming for all reports assigned Pho- tolysis, plus all of the reports assigned Ultraviolet radiation but not Photolysis, plus all of the reports assigned Photochemical reactions but not Photoly- sis or Ultraviolet radiation. The percentage of irrelevant information would be quite high." 282 The First Revision of the Thesaurus of ASTIA Descriptors, written bv Eugene Wall and issued by ASTIA August 6, 1962 [622] , is the final report on the contract between ASTIA and the Engineers Joint Council (EJC) . This report states that the purpose of the revision was "to create for ASTIA an improved indexing and retrieval authority, based upon Edition I of the Thesaurus, and to insure that this authority would be as useful as possible to organizations other than ASTIA." 283 The descriptors in the first edition were reviewed for deletion or addition, "for elimination of synon- ymous terms, for provision of extensive cross- referencing, for addition of scope notes to those descriptors needing such treatment, and for assign- ment of descriptors to subject-matter fields-of- interest. To achieve compatibility with other operating systems, it was necessary to call on the advice of dozens of other systems operators and subject-matter specialists . . . .' ,284 The greater portion of the second ASTIA Thesaurus is the scope note index, which lists the descriptors and use references in alphabetical order. The indexing terms or descriptors are again arranged into descriptor groups, now re- duced to 170 in number, which are themselves arranged into descriptor fields (increased in num- ber from 19 to 26) covering the AD collection in terms of subject matter disciplines. A new fea- ture is the display of generic relationships for 281 Hicks, 1963 [260], p. 145. 283 Hicks, 1963 [260], p. 146. *> Wall, 1962 [622],, p. 1. 284 Ibid. some descriptors in terms of "specific to" and "ge- neric to" entries. The ASTIA Chemical Thesaurus, a supplement to the Thesaurus of ASTIA Descriptors, second edition, has fragment descriptors and specific guidelines for indexing organic compounds. The fragment descriptors designate the functional groups and the substituted groups, radicals, or ions of the molecular structure of the compound. The reasoning of the editors was that compounds such as drugs, biologicals, enzymes, and vitamins should be indexed by these appropriate terms when their chemical structure was not under study. The in- troduction to the chemical thesaurus states that the system is intended to entail "coding the chemical structures in a given document into one code for- mat without the use of links, interfixes, or other devices for keeping the information separated. Thus, when chemical structures are composited, a new structure that contains all of the characteris- tics of the individual compounds appearing in the document is generated within the code sheet." ■« By November 15, 1963, eleven "microthesauri" had been compiled by ASTIA-DDC in the follow- ing subject areas: biological warfare, bionics, chemical rocket propulsion systems, masers, micro- biology, physical oceanography, psychophysiol- ogy, radiobiology, refractory metals, semiconduc- tor devices, and ultraviolet and infrared detection. These were compiled in order to allow more de- tailed categorization of information than per- mitted by the regular Thesaurus. In order to make these microthesauri have "general accepta- bility and compatibility" contributions and criti- cisms from experts outside ASTIA were requested. DDC is currently making possible (as of October 1963) rapid telephone search service in these dis- ciplines to organizations having an approved Field-of -Interest Register (FOIR) at DDC. An early effort by a professional organization to compile thesauri has made considerable impact. The first part of the Chemical Engineering The- saurus, published by the American Institute of Chemical Engineers (AIChE) in 1961 as "a word- book for use with the concept coordination system of information storage and retrieval" ( [18] ) , lists terms of a chemical engineering nature, and the second part lists purely chemical terms. The Chemical Engineering Thesaurus was de- veloped from one originally prepared by the En- gineering Dept. of E. I. du Pont de Nemours & Co. Each term of the thesaurus was reviewed for ge- neric relationships and semantics, as well as its suitability for inclusion, by a subcommittee of the AIChE Standards Committee. The types of references used under the terms are different. For example, since ethyl alcohol is not a term in the thesaurus, it is merely listed alpha- betically in the second part as Ethyl alcohol see Ethanol. Under Ethanol there are two notations : 2s 5 Defense Documentation Center, 1962 [162], p. 1. ro SF Ethyl alcohol (meaning that the person perus- ing Ethyl alcohol has been told to see (seen from) Ethanol) and PO Alcohols (meaning that Etha- nol has been "posted onto" the term Alcohols, thereby causing the latter to bear the notation GT Ethanol). This indicates that Alcohols are "ge- neric to" Ethanol and a great many other alcohols, that to be truly all inclusive any document having information on ethanol may be regarded as having information on alcohols as well. Another notation is RT, applied to "terms which have a similar meaning or a frequent association with the main term, but which are not strict generics or syno- nyms . . ." 286 meaning that they are "Related Terms." For example, Flameprooftng has two RT listings under it, Protection and Safety. Under Protection, similarly, there is the reference RT Flameprooftng but not RT Safety, while Safety has RT's to both Flameprooftng and Protection. These "related terms" are similar to "see also" terms. As partial explanatory text to the Chemical En- gineering Thesaurus there are included two articles reprinted from the May and June 1961 issues of the periodical Chemical Engineering Progress; Morse states that these issues "will form the AIChE Information Retrieval Standards . . . ." 287 Morse also illustrates the use of keywords and catalog cards in concept coordination and the ap- plication of links and roles. He states that "numerals (links) define groups of key words that were linked with one another in the original arti- cle. . . . The letters (roles) after each key word designate a code implying the way the key words were used — that is, the role of the key word in the original article," 288 for example, whether a specific chemical was used as a raw material, catalyst, in- termediate, or product. Holm in the June 1961 Chemical Engineering Progress gives a "how-to-do-it" explanation of the program initiated that month by the AIChE. The program was for retrieving technical literature in a publication where each technical article pub- lished was accompanied by both an abstract and key words supplied by the author. He considers that use of this system will be helpful to three types of users: the individual chemical engineer who needs a small personal file of minimum complex- ity; the medium-size organization with broader interests ; and the large organization serving many people with wide interests. 289 These activities of the AIChE led to the EJC sponsorship of a symposium in New York City in January 1962. The purpose of the symposium was to consider how the engineer might improve his ability to retrieve recorded information. The proceedings of that symposium contain addresses on (1) the size of the problem of storing and re- trieving technical information, (2) the technology ^American Institute of Chemical Engineers, 1961 [18], p. II. 287 Morse, 1961 [399], p. 58. 288 Ibid, p. 57-58. ""•Holm, 1961 [268], p. 76-77. used to attack the problem, (3) working systems for retrieval, and (4) a two-part "action plan" for the abstracting and indexing of technical arti- cles and books of lasting value concomitantly with their publication, and the creation of a single "Thesaurus of Engineering Terminology" for all branches of engineering. The "EJC Action Plan" stated that the cost for creating this thesaurus "will be relatively small because of the work al- ready done on the Chemical Engineering The- saurus and the steps already taken to insure com- patibility with the new ASTIA Thesaurus." 290 The EJC was given a grant by the National Science Foundation under which Eugene Wall, as Director of Information Services, conducted a "study of engineering terminology and relation- ships among engineering terms." The final report on the EJC study (August 1963, [186]) noted 18 contributors who responded to the request made to the national professional engineering societies and other organizations, "to submit any subject head- ing lists, thesauri, glossaries, or any other item lists which they employ as indexing authorities or which they consider useful for that pur- pose. . . . These eighteen contributors submitted about 119,000 terms, of which 87,550 were 'unique.' " 291 The EJC Engineering Terminology Study Com- mittee of 131 members, dividing itself into ten subcommittees, gave volunteer and detailed assist- ance in the research effort. "Twenty-seven sub- committee meetings were held, full-time for one week per meeting; 130 man-weeks of volunteer assistance were provided by the subcommit- tees ... A total of about 10,500 terms were treated during this period. A 1000 term sample of this proposed vocabulary was analyzed and from the results of the analysis, the following conclu- sions are drawn: (1) The investigational proce- dures used during the study were appropriate. (2) Overlap of term usage within the engineering prof essi on is significant (74%). (3)_ Ambiguities in term meanings within the engineering profession are fairly frequent (8%), but are easily resolved. (4) Most terms of any major degree of interest to even one portion of the engineering profession have been detected and considered. (5) It is feasible to provide a unified vocabulary for use by the entire engineering profession." 292 The Medical and Health Related Sciences The- saurus, issued in March 1963 by the Division of Research Grants of the National Institutes of Health, was developed during 1960-62 simultane- ously with the preparation of the initial annual Research Grants Index, a subject-matter index of research projects supported by the Public Health Service. This thesaurus has 12,200 terms and groups of related categories of information as con- venient broad access points for related concepts. ™ Ibid. p. 76. m Engineers Joint Council, 1963 [186], p. 2-3. JM Engineers Joint Council, 1963 [186], p. 1. 71 The groupings include (1) subheadings of three general kinds, subordinates of body systems or functional units of body systems, terms for com- monly accepted functional attributes, and taxo- nomic terms for collecting related information al- though the taxonomic arrangement is nonhierar- chial; (2) alphabetical groupings for "tied terms" of related data, for example, Radiolabeled arsenic and Radiolabeled carbon instead of the more exact Radio-arsenic labeled and Radio- carbon labeled, respectively ; and (3) indication by cross reference, rather than subheading, for the relationship of terms which would constitute an arbitrary grouping. For example, Epinephrins is made a main term with cross references to it from both Sympathomimetic agents and Catechola- mines, rather than having these terms grouped arbitrarily under drugs. The National Library of Medicine issued the second edition of Medical Subject Headings, as part of its January 1963 Index Medicus; the sub- ject headings include the main headings and cross references used in Index Medicus and the National Library of Medicine catalog. Although the vol- ume is not termed a thesaurus, it contains a total of some 5,700 "descriptors", Another listing of these descriptors by 13 categories is appended, the categorization of terms being intended "as an ordering of terms to suggest relationships, not as an official classification. . . ." 293 Each subcate- gory list is preceded by a group of major generic headings under which most of the terms within the category will fall ; they serve as guides to the materials in the list. Two thesauri were prepared in late 1963 by other U.S. Government agencies. The Bureau of Reclamation in Denver prepared the tentative edi- tion of its Thesaurus of Descriptors: A List of Keywords and Cross-references for Indexing and Retrieving the Literature of Water Resources De- velopment in October 1963. Camp ( [599] ) of the Technical Library of the Bureau of Ships com- pleted the compilation of a Thesaurus of Descrip- tive Terms and Code Booh for the Bureau in De- cember 1963. The Bureau of Reclamation Thesaurus of Descriptors is presented in three parts. Part I is an alphabetical list of descriptors and "use references", which constitute the main indexing and retrieval tool. Part II gives first a listing of 47 descriptor groups, arranged alpha- betically, with indication of that descriptor "field" to which each group belongs, and then gives a listing of the descriptors pertaining to each group. Part III lists the six descriptor fields and then the descriptor groups pertaining to each field. These fields are civil engineering; earth sciences and agriculture; electrical, mechanical, and industrial engineering; materials engineering; physical, chemical, and biological sciences; and social sci- ences and miscellaneous. The Thesaurus of the Bureau of Ships was de- veloped for use in conjunction with Project SHARP (SHip Analysis and Retrieval Project), a cooperative project between the Technical Library and the Applied Mathematics Laboratory of the David Taylor Model Basin for the design and testing of an automated information storage and retrieval system for the Library's collection of 170,000 reports, using the IBM 7090 and 1401 com- puter facilities at the Laboratory. This thesaurus is "a word-association list generically structured to enable indexers and subject analysts to describe the subject information of a document to the de- sired level of generality or specificity at input, and to permit searchers to describe in mutually precise terms the information required at output. It is a flexible authority list for vocabulary control with provision for the use of terms in combination, for concept coordination, in contrast to standard li- brary subject heading lists which do not permit this degree of flexibility." 294 The method of in- dexing of the EJC was used for "facilitating the formal manipulation of document descriptions and queries by computer." 295 Finally, we reemphasize the reported close kin- ship between the modern thesauri and the more traditional subject authority lists. Thus, "One of the keystones of the entire [MEDLARS] system is the subject heading authority list, Medical Sub- ject Headings . . . the 1st edition . . . was care- fully compiled in 1959 ... to prepare for the computerized retrieval system, the full scale re- vision of this list was initiated. The huge task of converting 4400 main headings and sixty-seven standard topical subheadings contained in the 1959 list into a 5700 main heading system with no topical subheadings was completed in October 1962 and published in January 1963. The cross-refer- ence structure was augmented by a completely categorized arrangement of the main headings in addition to the purely alphabetically ordered listing." 296 3.7.2. Classification Schemes The search for a systematic classification of all knowledge has had a venerable but erratic his- tory, reaching back at least to the Aristotelian categories. Vickery indicates that the "subsequent history of classification has been the attempt to find a new rational system of the sciences to re- place that of Aristotle. . . ." 297 He points out, however, that "history presents a series of cultural epochs. Each is a span of years within which knowledge presents a more or less unified struc- ture which can be expressed in a classification, but each new epoch requires a new classifica- tion . . . Indeed, the better fitted a classification 293 National Library of Medicine, 1963 [411]^ p. vi. 294 Bureau of Ships, 1963 [600], p. vi-vii. 295 Ibid, p. iii. 296 Taine, 1964 [552], p. 123. 297 Vickery, 1959 [610], p. 159. 72 is to a given epoch, the less suitable will it be for any other epoch." 298 Similarly, Shera suggests that : "Even a cursory examination of the history of the classification of the sciences emphasizes the extent to which any attempt to organize knowledge is conditioned by the social epistemology of the age in which it was produced. This dependence of classification theory upon the state of the sociology of knowl- edge will doubtless be even more strongly con- firmed in the future. . . ." 2 " Pings is among those who are convinced that universal classification is untenable because each science or discipline develops its own ways of un- derstanding the elements of that discipline, and that to generalize a method for covering all sciences or disciplines would lead to confusion of terminology and ideas. He states: "A standard- ized vocabulary can only be a universal classifica- tion scheme if both the subject content and the language are dead. It is quite possible that all the knowledge and events of the Roman Empire could be classified in Latin. The significant and erroneous knowledge of these centuries found its way into the language. The efforts of bibliogra- phers to standardize language only compounds the natural property of language to keep pace with the experiences of man." 30 ° Nevertheless, since at least the seventeenth cen- tury and earlier, general purpose or universal pur- pose classification schemes have been developed and applied to problems of bibliographic control and to the ordering of books and documents on shelves. Early examples are cited by Shera as follows : "Bacon's scheme influenced the early clas- sification system of the Bodleian Library. Thomas Jefferson founded upon it the classification for his own books and from this it was absorbed into the plan of book arrangement at the Library of Con- gress, where it was employed, with modifications, for almost a century. In inverted form it was used by William T. Harris, from whom Melvil Dewey took it for his own Decimal system. . . . The classification devised by Konrad Gesner for the organization of his Pandectarium sive parti- tionum universaliwm, [was] considered by Ed- wards to be the first bibliographic system, and certainly the greatest early attempt to relate the subject arrangement of books to the educational and scientific consensus of the day." 301 In this section we shall consider some of the major schemes that have been devised since the middle of the nineteenth century and that use a numeric or alphabetic notation for the classified organization of information. The decimal classification of Melvil Dewey, first published in 1876, has had a marked influence on schemes for classifying books and is used by most of the public libraries in this country. The 1951 208 Ibid, p. 158. 209 Shera, 1951 [5237, p. 82. 800 Pings, 1960 [4581, P. 12. 301 Shera, 1951 [523], p. 74, 75. 76. edition of the Dewey Decimal Classification, the 15th, was compiled within the Library of Con- gress, to be succeeded by the 16th edition pub- lished by the Lake Placid Club in 1958 (Dewey 1958 [174]). Besides its extensive use in public libraries, the Dewey Decimal Classification has served as the basis for the Universal Decimal Classification (UDC) which has become the widely used European classification for libraries and for denoting the subject categories of articles in scien- tific periodicals. Because of its acceptance in many areas as a common classification language and because it is has been established and is main- tained as an international cooperative effort, the UDC will be discussed in more detail below. First, however, we Avill consider briefly other ap- proaches to universal or general purpose classifi- cation systems. C. A. Cutter compiled a series of seven succes- sive classification schemes, each an expansion of the other. Cutter did not approve of the notation or order of Dewey's classification and worked out a new order which he claimed to be evolutionary within its divisions and which used for notation the letters of the alphabet and numbers one to nine. An outline was published in 1879. When the Ex- pansive Classification, Part I; The First Six Classifications, 1891-1893, appeared, it was de- signed for collections of every size "from the vil- lage library in its earliest stages to the national library with a million books." The LC classifi- cation was influenced considerably by Cutter's work. Cutter also published his Table, alphabetical order schemes consisting of decimal numbers so constructed that they may be combined with ini- tial letter or letters of surname or words, thereby supplying an alphabetical method for arranging items in a file and books on the shelves by the names of the author or originator. The Cutter numbers, or some variant thereof, are used exten- sively today. The best known of these variants is the Library of Congress Classification. "Starting in 1901, the classification system of the Library of Congress (LC) began to appear .... This is a system of traditional type ; it is entirely enumera- tive, and it would be useless to seek therein an expression of general categories: Specific subjects only will be found. Fundamentally, it is, after an interval of two centuries and a half, an under- taking very similar to that of the great librarians of the Bibliotheque Mazarine, and later those of the Royal Library in Paris, i.e., a careful and detailed arrangement of the greatest collection of books in the world, on the basis of groupings of the latter, let us say by affinities." 302 Landau records that "Dr. Herbert Putnam was appointed Librarian [of Congress] in 1899 and brought to fruition the gigantic task of arrang- ing a great national library according to a mod- ern classification. It is doubtful if such a project 303 de Grolier, 1962 [170], p. 43. 73 can ever be completed .... Thus, the [LC Clas- sification] scheme itself, while most scholarly in detail, does not pretend to have any but the loos- est connection between its main classes. . . . The schedules consist of a series of separate main classes each of which has its own relative index. There is no index to the whole, but the list of Subject Headings Used in the Dictionary Cata- logues of the Library of Congress can be a sub- stitute." 303 These limiting characteristics must be taken into account when considering schemes for classifying or indexing information for mechanized storage and retrieval of information. Similarly, in view of some of the problems and difficulties of melding central services with more specialized needs, it might be noted here that it was found ". . . impossible . . . for the Library of Congress to modify its medical classification sufficiently to persuade the Army Medical Li- brary ... to refrain from developing its own system. Probably both libraries were correct in their respective decisions . . . ," 304 _ Still another example of a traditional enumera- tive type of classification scheme is that of Bliss ([65]). _ "In 1910, Henry Evelyn Bliss published in the Library Journal 'A modern classification of libraries, with simple notation, mnemonics, and alternatives.' The 'Bibliographic Classification' . . . did not, however, begin to appear in abridged form until 1935 . . . Bliss developed the use of auxiliary tables (systematic auxiliary schedules) . . . He recognized also, in a certain measure, the inadequate character of a classification system in which the main classes are based on disciplines and not on subjects studied, and established numerous 'alternative locations,' so that the same complex science could be placed in several different classes ... It was only in 1948, perhaps under the influence of Ranganathan's ideas, that he advo- cated the use of relationship signs." 305 Faceted classification schemes, and especially the prototype Colon Classification of Rangana- than, are directed both to multi-aspects identifica- tion of subject content, reflecting different possible approaches, and to improvements in notation. Speaking to the latter point, Ranganathan himself has discussed limitations of prior notation schemes and the solution of rigidities as follows: ". . . the field of knowledge soon outwits even the clev- erest guess about the gap that should be left between any two integers brought into use at the time of designing the classification. It often hap- pens that certain gaps do not get filled up at all, while certain others soon get glutted and it is in these glutted regions that newer and newer specific subjects come up. . . . ". . . Rigidity was first broken by the simple device of numbering the known specific subjects, not by consecutive integers, but by nonconsecutive ones leaving a reasonable number of unused integers between them — by gap notation, so to speak. . . . "The greatest forward step in classification was taken less than a century ago by one whom I always call the 'Father of Modern Librarianship' — Melvil Dewey. He broke the second rigidity in notation ... by abandoning the use of integers altogether and introducing pure decimal-fraction-notation . . . [which] provides for infinite hospitality." 306 Ranganathan points out further, however, that "hospitality of notation has to be two-fold — in chain and in array" and that ". . . in a faceted notation — in the Colon Classification — a class number can grow not only at its end but also at the end of each of its facets." 307 The multi-aspect subject approach of Colon and faceted classification has been summarized by Vickery as follows : "The first explicit use of cat- egories in classification was made by Ranganathan. His Colon Classification exhibited from the start a division of classes into facets or categories. For example, in his main class Medicine, before start- ing a family tree subdivision, Ranganathan first formed two groups of terms: organ (e.g. femur, diaphragm, blood, eye, tooth) and '■'problem' 1 (e.g., morphology, physiology, disease, hygiene). . . . "Ranganathan also found the need to intro- duce other groups of terms, to express the symp- toms of disease (e.g., fever, inflammation), its agents (e.g., virus, bacteria, poison) and its handling (e.g., surgery, diet, nursing, therapy). He was thus able to form compound headings of the type FEMUR— MARROW INFLAMMA- TION (OSTEOMYELITIS)— PHARMACO- THERAPY .... The Colon Classification, by dividing its terms into categories or facets, and making notational provision for linking facets together, was thus able ... to bring together terms [in coordinate relationships]. By prescrib- ing a fixed combination order for the facets, the classification settled the problems which worried Cutter and Miss Prevost." 308 Other remarks may be noted here which are related to the notion of universal classification schemes. The possibility of a systematic classifi- cation of all of knowledge, of course, has been an enticing activity as long as there has been re- corded knowledge at all. Ball suggested in 1947 a "National Library of Science, patterned after the present Library of the Department of Agri- culture, with facilities organized on a scale to em- brace all of the fields of scientific and technical effort. ... A system of classification such as is used in the Patent Office, integrated perhaps with the Universal Decimal System, could serve as a basic framework on which to build a really effec- tive system of technical and scientific classifica- tion. Such a system would justify a corps of 303 Landau, 1958 [327], p. 189. Note however, that an index to each volume is provided. a" Evans, 1963 [188], p. 15. *° s de Grolier, 1962 [170], p. 43. 303 Ranganathan, 1951 [468]i, p. 98. 807 Ranganathan, 1951 [468], p. 100. 3 <» Vickery, 1959 [610], p. 12-13. 74 specialists continually supervising its development in accordance with the needs of scientific and tech- nical workers using it. A council of recognized leaders in the major fields of science and tech- nology should meet at periodic intervals to study proposed revisions and guide the efforts of the staff to maintain the system effective and accepta- ble to all concerned." S09 Newman followed with the suggestion that the classified files of the U.S. Patent Office could be used as a basis for the classification of technology. He suggests that the existing system might be en- larged to include theoretical science (1961 [420]). He noted later that "Classification experts believe that the best classifications for documentation are those which classify differences. . . . Since each patent contains one or more claims which precisely define the advance made by the inventor in rela- tion to that which was heretofore known, the ability to create an hierarchical classification by the differences between claims is possible. Fur- thermore, the subclasses of this classification often utilize relationships between concepts to set out these differences. It is clearly demonstrable that, when looking for related information, the specific relationship of the elements of a 'thing' is many times more important and fruitful than the spe- cific details of the elements so related." 81 ° These characteristics of the Patent Office classification scheme permit dynamic revisions in meeting changing needs and emphases. Wahlin of Sweden (1963 [614] ) proposed a uni- versal system of classification based on funda- mental concepts arranged with consideration to their generic relations. His Ideal Universal Sys- tem (IUS) has such groupings as abstract basic concepts (number, space, time, motion, force, mass) ; energy and matter; the universe and the earth; physical life; the individual; society; ma- terial civilization; intellectual culture; and coun- tries and people. These concepts are those com- monly given as the "subjects" of the various fields of knowledge. He has also developed a variant system, the Technical Universal System (TUS) that is adapted to engineering and technology. "Here those branches of engineering and technol- ogy which constitute applications of some elemen- tary fundamental concepts (e.g., electrotechnics) are combined with these fundamental concepts, theory and practice keeping each other company to an increasingly low level the more specialized the concept and its application are." 3U Foskett reports that the Classification Research Group of England has met nearly every month since 1952 "with no resources beyond the native wit of its members, no allegiance to any existing system of classification, no fixed target . . . ." The group published a memorandum in July 1955, concluding "that any subject index, alphabetical, 309 Ball, 1947 r44],p. 35. 310 Newman, 1901 T424], p. 208. 311 Wahlin, 19G3 [619], p. 185-186. classified or mechanical, relies for its efficiency on the coordination of concepts or ideas expressed in documents, so that a set of concepts stated as the requirement of a research worker can be matched with the same or similar sets already .existing in the library or information system." 312 Foskett further states that although the Group rejected all existing schemes it "decided to adopt at least some of the Colon terminology and method" of S. R. Ranganathan. The Group adopted the term " 'facet' ... to describe the hom- ogeneous groups of terms that result from divid- ing a subject into its appropriate categories. It did not seem, however, that Ranganathan's princi- ple of deriving the facets of a subject by their relation to five 'fundamental' categories (Person- ality, Matter, Energy, Space, and Time) offered v a wholly reliable extension of the idea of cate- gories, and the CRG has never accepted this as a general theory . . . ." 313 Turning now to the Universal Decimal Classi- fication, we note first that it was developed by the International Institute of Bibliography (estab- lished in 1895, later to become the FID). Under the direction of Paul Otlet and Henri LaFontaine, the objective was to develop for purposes of bibli- ographical control a comprehensive word repertory for the organization of knowledge. "After ex- amining existing systems, Otlet and LaFontaine concluded that the Dewey Decimal Classification . . . offered the most promising basis. . . . Per- mission to modify and expand his scheme was ob- tained from Dewey on condition that the order of main classes and divisions be maintained and that maximum compatibility in development be sought. . . . "The new scheme, the Classification Decimale (CD), was first published in a complete interna- tional edition (in French) in 1905 as the 'Manuel du repertoire bibliographique universel.' From 1927 onwards, further full international editions, embodying the fruits of international cooperation in additions and revisions, appeared in French, German, and English together with numerous abridged editions in these and other languages." 314 While UDC has been widely accepted abroad, with a few notable exceptions it has not been as generally received or used in the United States. The collection of the Engineering Societies Li- brary is classified according to UDC, somewhat modified. G. K. Hall & Co. of Boston announced in late 1963 the publication in 13 volumes of the classed subject catalog of the library. The direc- tor of the library, Ralph H. Phelps states in the announcement that the library with its 180,000 volumes "acts both as an archive for the older material and as a working library of current in- formation, primarily for the graduate and prac- ticing engineer, but also for the engineering stu- dent, as well as other persons needing technical 312 Foskett, 1962 [199], p. 127. 313 Ibid. p. 128. »" Mills, 1964 [388]), p. 9. 75 information." The published catalog has 212,400 cards and its index 27,300 cards. Additionally, the John Crerar Library, the Li- brary of the U.S. Naval Ordnance Test Station at China Lake, the Technology Department of the Carnegie Library of Pittsburgh, the U.S. Weather Bureau Library, the Library of the Housing and Home Finance Agency, the Columbus Memorial Library in Washington, and (understandably) the United Nations Library in New York are among the organizations that use the UDC classification scheme. The general introduction to the abridged Eng- lish edition of the Universal Decimal Classifica- tion (UDC), published by the British Standards Institution (1961 [82]), states that the UDC "is a scheme for classifying the whole field of knowl- edge." Further, there are three basic principles involved in its construction and use: (1) It is a classification, "depending on the analysis of idea content, so that related concepts and groups of concepts are brought together. . . ." (2) It is a universal classification where the "universality at the conceptual level is supported by notational de- vices, which permit the linking of simple main numbers (for simple ideas) either with other main numbers or with auxiliaries denoting Place, Time, and similar commonly recurring categories — in each case forming combined or compound num- bers." (3) It is a universal decimal classification "proceeding from the general to the more particu- lar by the (arbitrary) division of the whole of human knowledge into ten main branches, each further subdivided decimally to the required degree." The first volume of a new series on the Intel- lectual Organization of Information being pre- pared under the aegis of the Rutgers Library School (Mills, 1964 [388]) provides a comprehen- sive study of the UDC, clarifies many of the com- mon misconceptions about it, and highlights both advantages and disadvantages of such a universal classification scheme. Mills emphasizes that while the UDC is usually employed in systems involv- ing conventional card catalogs or book-form in- dexes it can be used in term-entry as well as item- entry systems 315 and that it can be adapted to mechanization in several ways: "(1) It provides a carefully organized and very comprehensive vo- cabulary of terms, providing a search program- ming aid (by indicating a very large range of connections) more thorough than any existing thesaurus. (2) The terms are the product of in- tensive analysis, they exist in a form sufficiently elementary to allow a wide range of post-coordi- nation ... (3) Its notation is often hierarchical and is readily usable as a code (for punched-cards, say) expressing hierarchical relations . . (4) The production of full, detailed schedules, continually 315 Mills, 1964 [388], p. 54. 76 updated, and perhaps in a number of different languages, is susceptible to mechanization. . . ." 316 Among the advantages of use of the UDC cited by Mills are that it has the widespread acceptance of more than 5,000 organizations, that it can be applied both to books and to materials other than books, and that it has a highly organized machin- ery for revision, maintenance, and development, with extensions and corrections issued at regular 6-month intervals and cumulated every 3 years. 317 On the other hand: "The physical difficulties of altering documents and indexes is a serious prob- lem . . . and a major problem in maintaining UDC is that of keeping a balance between conser- vation in the retention of existing locations (favored by older users with substantial files already established) and the pressure of new mem- bers for relatively radical changes. There is a formal 'ten-year-rule' whereby a number which is cancelled is not used with a different meaning until 10 years have elapsed. Some users regard this as unnecessarily restrictive and it is some- times ignored in private files." 318 Tackling some of the common misconceptions about a scheme such as the UDC, Mills first points out that it need not be lacking in specificity of subject content indication down to the 'micro- thought' level as required. "... A single UDC number may represent several different terms; e.g., 543.361.3 is equivalent to using the following terms : Chemistry, Chemical Analysis, Analytical, Water, Salt Content, Halides, Fluorides." 319 Secondly, the UDC type of system need not be limited to use of the classified structure alone but is effectively supplemented by an alphabetized ("A/Z") index to the classification schedules. Such an index performs several major f mictions: "It is a key to the location of classes, providing a lead-in from the terms of the natural language to the highly artificial notational language. ... It remedies to a significant degree the great draw- back . . . [of a classified index] the scattering of what are 'distributed relatives' .... All the dif- ferent classes under which the subject is scattered are automatically assembled together as qualifiers of the common factor. . . . The ambiguities of language and common usage will always be a cen- tral problem in information retrieval; a classifi- cation like UDC remedies them partly by referring things to systematic contexts. If the reader asks for 'Rockets' but really means 'Missiles' he will soon find this out when he finds the context is one of engines only and not vehicles, and the A/Z index often provides this sort of 'classification' by its qualifying terms." 320 Examples of adverse reactions to the UDC to be found in the literature typically include the com- plaint that the UDC is not continuously revised, that although "many committees and single con- 310 Ibid, p. 14. 317 Mills, 1964 [388], p. 11, 14. 318 Ibid, p. 68. 3,9 Ibid, p. 85. 320 Mills. 1964 [388]-, p. 25-26, 105. tributors are trying to keep the UDC up-to-date . . . tens of thousands of UDC users either do not know of these changes and innovations or do not care to insert them into the tables and to classify accordingly . . . ." 321 Wellisch proposes that the user of the UDC subscribe only to those parts of the tables that he needs and that there be an adequate index. Vickery, in an article on the UDC and technical information indexing (1961 [469]), makes such suggestions as having schedules for new subjects such as plasma physics and solid state physics, re- consideration of notation in order to achieve shorter class numbers, and reorganization of the physics and physical chemistry sections. Vickery considers "that UDC notation is not useful for mechanized search systems, although the schedules themselves provide a valuable source of terms and hierarchies for such systems." 322 In contrast to Vickery's opinion with respect to search systems, the situation with respect to an- nouncement bulletins and index preparation is one indicating the feasibility of mechanization. Mills reports consideration by the U.S. National Com- mittee for the F.I.D. of machine production of UDC schedules in different languages and, further, that "recently a few examples of UDC indexes on tape have been reported, although these are primarily for the production by computer printout of book-form indexes for scanning in the normal way." 323 The Meteorological and Geoastrophysical Titles, second experimental issue, is an example in being. Rigby, in the introduction to this issue, reports on the success of experiments to mechanize the UDC, undertaken under a grant from the National Science Foundation. Some pertinent comments are "Owing to the varying length of the decimal groups, and to a number of special features such as brackets, colons, diagonal lines, etc., many doc- umentalists have concluded that the UDC was not 'machineable' and that it should even be dis- carded as an internationally recommended system for classification. Our experiments during the past year to use this system for arrangement and cross-indexing of titles, have, as can be seen from this issue, led to just the opposite conclusion, namely that the system is ideal for systematic in- dexing in almost any field of knowledge requiring fairly simple machines, and that with more so- phisticated machines such as the IBM 1401 or 704 used in the KWIC Indexes, or their equivalents of other makes, a superb job of organizing material of any complexity at any 'depth' could be achieved with a little experimenting." 324 It has been further reported that these experi- ments have "demonstrated that the UDC scheme is suitable for treatment with tabulators or elec- tric computers, and that in spite of some small difficulties the advantages, and possibilities for enhancement, of the use of UDC are greater than had been expected." 325 3.7.3. Interchange of Indexing Languages Five studies have reviewed the possibility of the interchange of indexing terminology. The first one, the Datatrol study of Hammond and Rosenberg in 1962, considers the questions of pos- sible convertibility between the vocabularies of several agencies ( [236] ) of the U.S. Government. A related report of Datatrol, by Jaster, appeared in February 1963 ( [296] ) . The third study, that of Painter in 1963 ([445]), is concerned with the duplication and consistency of subject indexing involved in report handling at the Office of Tech- nical Services. A fourth study of a large-scale transformation of one of the basic documents men- tioned in the other studies, although not about in- terchange as such, is the work of the Information and Documentation Center of EURATOM, Brus- sels, Belgium. This study is described in the re- port by Rolling, A Keyword List for Machine Documentation in the Nuclear Field (1962 [491] ) . The fifth report is Datarol's Technical Report IR- 10 of December 1963 ([155]) on the question of common vocabulary approaches for government scientific and technical information services. In the second half of 1961, the Ad Hoc Inter- agency Study Group on Language Compatibility in Mechanized Storage and Retrieval Systems was first organized in Washington. The formation of this group was an indication of Government rec- ognition for need of greater cooperation and com- patibility to avoid unnecessary duplication of ef- fort. Although there was no attempt at first to define "compatibility" accurately, it was suggested that a precedent had already been set by the Group for the Standardization of Information Services. Gray had pointed out that the earlier Group had found enough common subject headings "so that the subject heading problem will not prevent the interfiling of cards from the various agencies and the termination of duplicate processing." 326 and Taube had the opinion that "some progress has been made towards the development of a single standard list although it must be reemphasized that the final achievement of a single list remains a future contingency." 327 What is now desired is a degree of compatibility permitting the interchange of items for biblio- graphic control contained on such recording media as magnetic tapes, not only within such an orga- nization as NASA (having research establish- ments in various parts of the country) but also among NASA and like members of the govern- ment "reporting community" (AEC, DDC, and the Clearinghouse of Federal Scientific and Tech- 321 Wellisch, 1060 rfi.301, p. 145. 322 Vickerv. 1961 [616], p. 136. 823 Mills. 1964 [3RS1. p. 43. 824 American Meteorological Society, 1962 [20], p. 1. 325 As reported in Scientific Information Notes 5, (August- September 1963), p. 9. 320 Gray, 1951 r2211, p. 8-9. 827 Taube. 1952 [560], p. 23. 77 nical Information) and with the National Library of Medicine and the National Agricultural Li- brary. Hammond suggested in August 1962 that "a hypothesis has been advanced that converti- bility is a more attainable objective, particularly in view of the assistance now available from mod- ern computers." 327 In September 1962 Hammond was differentiat- ing between compatibility and convertibility, not- ing that the latter is "the ability to go from one indexing vocabulary to another, defining the path in such a way that any incompatibility is circum- vented. For example, in that the ASTIA descrip- tors and AEC subject headings cannot be inter- filed with a resulting one-to-one match, and in that a composite of the two vocabularies would be excessive for either agency, the two vocabularies would be called 'incompatible.' Since, however, equivalents — that is terms that one or the other would use as substitutes for the others — can be determined, the vocabularies are 'convertible.' " 328 Painter's study distinguishes between converti- bility and compatibility as follows : "Convertibil- ity means the ability to take separate operations, or, in this case, information, draw parallels or equivalencies, and then through some type of manipulation be able to extract similar informa- tion from all units; convertibility implies similar intellectual expression. Compatibility, on the other hand, means that while the separate units are completely dissimilar as far as both system and intellectual expression are concerned, the out- put of the system can be used by another system and the information supplied can be retrieved. There does not have to be equivalency at all. Compatibility is usually applied to systems and programs or physical objects, while converti- bility applies to terminology, information or in- tellectual expression." 329 Can the compatibility of indexing terminologies be achieved by having them convertible? The seeds to a first approximation for resolving the problems of interchange of such terminologies among the scientific agencies of the U.S. Govern- ment might be found in the answer to this ques- tion. In the Datatrol studies, the essential task was to take each of the 7,145 descriptors in the 1960 first edition of the Thesaurus of ASTIA descrip- tors and find its identical, synonymous, or use- fully equivalent counterpart in the third (Janu- ary 1960) edition of Subject Headings Used by the US AEC Technical Information Service. The ob- ject was to develop if possible a table of indexing equivalents, which would be unidirectional in that it shows AEC equivalents for ASTIA's vocabu- lary, but not vice versa. Datarol "decided to ap- proach the task in ASTIA's Group and Field en- vironment. While one-for-one comparison in straight alphabetic order would reveal identical 327 Hammond and Rosenborg, 1962 [236], p. 1. 328 Hammond, 1962 [235]i, p. 224. 329 Painter, 1963 [445], p. 99. terms, it would be inefficient, more difficult, and less meaningful for finding other relationships." 330 The 292 descriptor groups of ASTIA's first Thesaurus had been compartmentalized in order to gain coherency within relatively small group- ings of descriptors. The 292 groups were then fur- ther placed into 19 fields. The 7,145 descriptors are subsumed under the 292 groups, and the latter are subsumed under 19 broad fields. Datatrol took each one of the 13,385 subject headings of the AEC and subsumed it under (assigned it to) one of the ASTIA groups, a not inconsiderable task. "Each ASTIA Descriptor was analyzed individu- ally in its Group and Field environment and in context with related AEC terminology. Eight separate degrees of convertibility were established so that the ASTIA terms could be categorized as to their degree of convertibility for closer study. A numeric code was assigned to each category." 331 These eight degrees of convertibility between descriptor and subject heading were as follows: (1) identical terminology in spelling and con- text; (2) synonymous but not identical terminol- ogy; _ (3) standardized AEC subheading 332 is identical to, or synonymous with, the descriptor; (4) a subject heading in combination with a sub- heading is synonymous with or generic to a de- scriptor; with the dropping of the subheading the relationship would be represented by No. 6; (5) two or more subject headings are required for equivalent indexing terminology; (6) a subject heading of a broader generic level subsumes the descriptor; (7) the descriptor of a broader generic level subsumes the related AEC terminology; and (8) no equivalent AEC terminology could be iden- tified at any useful generic level. So far, then, Datatrol placed the subject headings of the AEC into the descriptor groups of ASTIA, associated specific subject headings with specific descriptors, determined that there were eight categories of convertibility, and indicated the particular cate- gory for each descriptor and the pertinent subject headings allied therewith. The bulk of the Datatrol report consists of its table of indexing equivalents thus developed. In this list, each descriptor from the 1960 ASTIA Thesaurus is arranged by ASTIA field and within each field by group : to each descriptor there are added (1) the AEC equivalent term or terms (if an y)> (2) the category of equivalence as deter- mined by Datatrol, (3) the ASTIA frequency of use (so far as Datatrol could determine it), and 330 Hammond and Rosenborg, 1962 [2361 p 3 331 Ibid, p. 4. y 332 The AEC subheading, standardized in January 1960 but no longer used as such, creates a problem in determining a final table of indexing equivalents. The current (1963) practice of AEC is to replace such terms by modifiers or key sentences under the entry for the main subject heading in the Nuclear Science Abstracts (NSA). The corresponding ASTIA terms are used analogously, namely "as modifiers through the technique of co- ordination, but they would not be used as main entry points in ASTIA s cumulative indexes. The terms in this category are represented by analysis, design, effectiveness, measurement, theory, and tests; the total number represents 2.0 percent of the ASTIA vocabulary, having been used almost 100.000 times oorg 196 y 2 ea [236] 7 \ A docum entation." (Hammond and Rosen- 78 (4) the code number assigned by ASTIA to that particular descriptor. For 38.1 percent of the ASTIA descriptors, the closest equivalent was an AEC subject heading of broader generic level; for 7.4 percent of the de- scriptors, the ASTIA terminology was of broader generic level than its closest AEC equivalent. The usage of ASTIA terms in both these groups for eight years of documentation represented only 29.7 percent of ASTIA's total usage of all terms. Synonymous terminology with AEC subject head- ings was found in 4.2 percent of the ASTIA as- signment to its collection for the eight year period. For 10.9 percent of the ASTIA descriptors there was no AEC equivalent, this percentage repre- senting only 6.1 percent of ASTIA's eight-year assignments. Datatrol's explanation for this last phenomenon is that in most instances 'AEC did not have a requirement for the ASTIA indexing terminology simply because it did not have docu- mentation of the same nature to index." 333 Painter reports that 85 percent of the AEC terms studied by Datatrol fell into three ASTIA fields, those for chemistry, materials-metals, and physics-mathematics, and that in these same three fields only 31 percent of the ASTIA descriptors appeared. Painter makes a distinction as to what may be inferred from the Datatrol study and the duplicate indexing analyses of her study on OTS : the Datatrol study "was done on a theoretical basis in the sense that it employs the total number of descriptors in the Thesaurus and the total num- ber of subject headings in the AEC list. . . . The Datatrol report attempts to show what is possible ; the duplicate indexing analyses check what actu- ally occurs in practice." 334 Datatrol insists that although its study should be considered experimental only, the "results indi- cate that patterns of conversion exist such that the ultimate goal of a Dictionary of indexing equivalents can be attained." 335 Further, the levels of convertibility are generally in the cate- gories of identity, synonymity, and generics. "Difficulties in the two latter areas exist, but are not insurmountable. The assigning, for example, of a higher generic term as a useful equivalent tends to be subject to nonrigorous criteria. Knowledge of the system users' requirements, however, can provide fairly realistic solutions. Other criteria can be devised as well." 336 Hammond speaking in September 1962 ( [235] ) , cited a specific instance of possible convertibility where there could be the substitution of an "equiv- alent" — an instance which may be regarded as un- fortunate in the light of subsequent events. De- scriptor group 98 on Food, of the first edition of the ASTIA Thesaurus, had a descriptor preserva- tion (which was unquestionably placed there be- cause the thinking in ASTIA was that most of such reports in its collection had been on jireservation of food) for which there was no equivalent AEC subject heading. Therefore, Datatrol determined to make the subject heading Food equivalent to ASTIA's descriptor Preservation (indicating, by the way, that this is a term of higher generic level). "In AEC's five-year cumulative index to NSA we were able to see that, in fact, AEC had placed documents on food preservation under the subject heading Food." 337 The dangers inherent in such subsumption of one entity under another are illustrated by the lo- cation of Preservation in the 1962 ASTIA The- saurus. In the revision, descriptor group 64, Food, contains only "things" (actual physical entities), for example, foods or collection of foods, as in dehydrated foods and military rations, while Preservation is placed with descriptor group 169 on Abstract Concepts. The entry for Preserva- tion in the revised Thesaurus itself has the term placed in descriptor group 179 on Industrial and Laboratory Processes (in which group, inciden- tally, it is not found) and the term is denned with a scope note as "Prevention of deterioration of stored commodities, structural members, mate- rials, etc." The descriptor Preservation has an "also see" in the thesaurus to Storage, which de- scriptor is placed in descriptor group 38 for Con- tainers and Packaging, this descriptor being the only one in group 38 that is not a "thing." Hammond also describes (1962 [235]) two studies which he contends support the earlier Datatrol study : convertibility of most frequently used AEC terms, and comparative analysis of the AEC and ASTIA indexing of a common body of reports. The first of these supporting studies had been reported in the first Datatrol study. It indicated that "the 905 most frequently used AEC headings (used at least once a month), while rep- resenting only 7 percent of the total AEC five- year vocabulary, accounted for some 80 percent of AEC term assignments. Over half of these headings were identical or directly convertible, and the others easily convertible to ASTIA de- scriptors. These figures clearly show that, to give a true picture, any measures of convertibility of terms must take into account their frequency of use, since each assignment represents an event hav- ing convertibility potential." 33S In the second supplementary Datatrol study, Jaster compared the AEC and ASTIA indexing of 277 reports cataloged by both agencies. The sample was selected from the AD reports listed in the Semiannual Report Number Index to NSA for the period January-June 1961. To index these 277 reports ASTIA used 2,571 descriptors and AEC 840 subject headings, about half of the latter figure being identical, or practically so, with the ASTIA assignments. In 59 of the documents, representing 21 percent of the sample, all AEC 333 Hammond and Rosenborg, 1962 [236];, p. 7. 331 Painter, 1963 [4451, T>. 67 and 94. 335 Hammond and Rosenborg, 1962 [236], Abstract. 330 Ibid, p. 19. '"Hammond, 1962 [235], p. 224. 333 Hammond, 1962 [235], p. 226. 79 terms were matched by ASTIA terms. Hammond concludes ". . . it was apparent that conversion of the ASTIA descriptors by means of the table of indexing equivalents would, in fact, substan- tially have generated the terms AEC used." 339 The Datatrol report by Jaster (1963 [296]) describes a tool devised for compiling a dictionary showing multilateral equivalence among the in- dexing terms of the AEC, DDC, and NASA. In the preparation of the "subsumption" scheme, the field and group structure of the ASTIA 1960 Thesaurus was modified by three techniques: de- letion, addition, and distribution of terms. There resulted 20 fields instead of the 19 of the ASTIA Thesaurus. One variance in procedure adopted in the modification of the ASTIA/DDC scheme was provision for assigning indexing terms to any number of indexing groups whereas in the prepa- ration of the first ASTIA Thesaurus each descrip- tor was assigned to only one descriptor group. In the preparation of this subsumption scheme consideration was given to the use of broad cate- fories, possibly to be attained by (1) merging the elds of ASTIA/DDC, the categories of the NASA Scientific and Technical Aerospace Re- ports, and the categories of the AEC Nuclear Science Abstracts or (2) extrapolation from other schemes for categorization or classification such as the AEC evaluation lists, ASTIA Distribution Guide, or NASA distribution categories. Datatrol continued this line of attack in its December 1963 report on common vocabulary ap- proaches for government scientific and technical information services. This study was given im- petus by the interest of the Committee on Scien- tific Information of the Federal Council for Science and Technology, and its object was "to develop and document guidelines and criteria to be used in reaching a decision on a single Gov- ernment-wide approach to an information re- trieval vocabulary." 340 Datatrol was assisted in its inquiry by an Interagency Task Group on Vo- cabulary Study, formed to act as expert consul- tants to COSI, particularly in matters regarding common vocabulary. The members of the Task Group, with a representative of the National Science Foundation as Chairman, included repre- sentatives from the AEC, DDC, NASA, National Agricultural Library, National Bureau of Stand- ards, National Library of Medicine, and OTS. At the invitation of the Chairman of the Oper- ating Committee of COSI, members of the Science Information Exchange participated in the study, and members of the COSI Operating Committee and other responsible and knowledgeable individ- uals were consulted. The three specific require- ments of the study were (1) to identify the alternative courses of action available for govern- ment-wide adoption respecting common vocabu- lary approaches for information storage, search, "'Hammond. 1962 [235], p. 227. 340 Datatrol Corp., 1963 [155], Foreword. and retrieval; (2) to evaluate each course regard- ing its economic impact on each agency within the COSI Operating Committee; and (3) to report on evaluations and on the guidelines and criteria use- ful for reaching a decision on a single, government- wide approach to a vocabulary. The following five approaches to the alternative courses of action of the first requirements were given consideration: (1) single thesaurus: prep- aration of one vocabulary reference for indexing all government scientific and technical documents ; (2) single subject heading list: preparation of one list of subject headings for indexing all govern- ment scientific and technical reports; (3) cor- related thesauri: preparation of a thesaurus by each agency for its use and preparation of a ma- chine-independent computer program for correlat- ing all government thesauri; (4) combined the- saurus and subject heading list: preparation of a thesaurus or list of subject headings by each agency "according to its option (status quo)," i.e., its current method of conducting its documenta- tion, especially its current tackling of vocabulary, plus preparation of one or more machine-inde- pendent programs for locating all words used by government agencies to describe identical objects and concepts; and (5) common subsumption scheme: "preparation of a single list of broad generic subject categories that will subsume the composite subject content of government research reports. Such a scheme to be used by each agency to categorize reports independently of the index- ing vocabulary. Preparation of a computerized application for correlation of indexing data among the agencies." 341 Datatrol elected to evaluate these five approaches on the basis, of effectiveness and cost of indexing for announcement and retrieval of the government scientific and technical information systems. Re- garding the indexing of the reports for purposes of their announcement, Datatrol concluded that the indexing by either AEC, NASA, or DDC did not satisfy the requirements of each other. Regard- ing the indexing of the reports for purposes of their retrieval, Datatrol saw "no evidence that, under the current modus operandi, a single the- saurus, subject heading list, or correlated thesauri would offer any significant advantage over the status quo for searching multiple collections. Superimposing a common subsumption scheme . . . appears to offer the most promising approach under either a computer-correlated approach or for mutual organization of indexing data of sev- eral agencies." 342 Furthermore, adoption of the common sub- sumption scheme by all operating agencies was regarded as possibly providing the points of de- parture for computer generation of indexing equivalents for searching different systems. Data- trol defines its common subsumption scheme as "a 341 Datatrol Corp., 1963 [155], p. 8. 842 Ibid, p. 72. 80 tool of convenience for grouping and displaying associated terms of a given indexing vocabulary to serve as an authority for indexing, storing, and retrieving reports or data from a given informa- tion system. The associations may be vocabulary oriented and/or agency program oriented . . . i.e., based on subject category, scientific discipline, project or other special interest, or a combination of any of these associations." 343 Datatrol thus regards the statistical association of indexing terms as the most promising ap- proach for correlating the indexing data of the different operating agencies for retrieval. 344 The multidimensional correlations of statistical asso- ciation techniques include the term profile or sta- tistically associated terms from the indexing data of a given system and the identification of equiv- alent search terminology employing association profiles of terms used by different systems to index reports of related subject categories. 345 The study by Painter, submitted as a doctoral thesis to the Graduate School of Rutgers, The State University of New Jersey, was conducted as part of a cooperative Office of Technical Serv- ices-National Bureau of Standards project exam- ining the feasibility of automation of information handling operations in OTS. The three primary objects of the Painter study were to determine (1) the types, amount and frequency of duplica- tion among reports submitted to OTS from other government agencies; (2) the amount and fre- quency, as well as the significance, of duplicate indexing efforts by the contributing agencies ; and (3) the consistency of subject indexing within major agencies contributing to OTS (DDC, AEC, and NASA) and also at the National Agri- cultural Library, which does not contribute to OTS. Painter summarized her findings regarding duplication of reports sent to OTS as follows: "Duplication, examined by the regular search procedure in practice at OTS, was found to occur at three levels: one necessitated almost complete processing before search, one required only partial analysis, and one needed only one search item for identification. The duplication rate occurred at a maximum of 50 percent with the second and third of these levels. Most of the duplication Avas iden- tified through the report or series number or the AD number. It was concluded, therefore, that machine search [for duplicates] was feasible at the simplest level of processing (report number or series number). This would put preparation time at a minimum and eliminate the duplicated materials." 345a Painter also undertook a determination of the "equivalency of terms and their possible convert- ibility as revealed in an analysis of the indexing of a report or reports by two or more agencies." 343 Datatrol Corp., 1963 [155], p. 53. '"Ibid, p. 71. 315 Ibid, p. 98-99. s-isa painter, 1963 [445], p. vii-viii. She took as basic the Datatrol study on the con- vertibility between the ASTIA descriptor and the AEC subject heading, and she made a comparison with the results obtained by Jaster of Datatrol, reported above. Painter examined 200 items which had been indexed by both ASTIA and AEC in order to learn what equivalency of assignment had been made by the respective agencies. She found that of the 2,217 ASTIA descriptors and 640 AEC subject headings there was a total of 426 terms that were either equivalent, word-by- word, or synonymous. "Thus the percentage of equivalency within the ASTIA. descriptors was 20 percent as opposed to 67 percent within the AEC subject headings. On the basis of the total num- ber of terms assigned (2,857), the degree of simi- larity in indexing between the two systems would be 30 percent." 346 This last figure was obtained as a ratio of the sum of the ASTIA equivalent terms plus the AEC equivalent terms to the sum of all ASTIA terms plus all AEC terms. Painter's summation of the two Datatrol studies and her own related one is that: "Taken as a whole, the three studies seem to point out a degree of convertibility between ASTIA and AEC. A large percent of the AEC documents could be re- trieved by converting to ASTIA descriptors, at least in the three fields where comparison is most valid. A third of this is exact equivalency, or synonymity; a majority of the comparisons of terms are complicated by generic (term in one scheme subsumes two or more of another) rela- tionships, but according to the basic Datatrol report this need not be a serious block to retriev- ability. Only 10 percent possibly would not be retrieved." 347 Painter suggests, however, that more evidence is needed that satisfactory retrieval results from the use of a term of higher generality. A comparison and analysis was also made for triplicate indexing between ASTIA, AEC, and NASA for a small sample of 16 items. To these 16 items, ASTIA had assigned 219 terms, NASA 74, and AEC 32. In these assignments there were 25 equivalents appearing among all three, making 12 percent equivalency with the terms from ASTIA, 64 percent from NASA, and 78 percent with those from AEC. Painter made two comparisons between assign- ments by ASTIA and NASA for 90 NASA docu- ments indexed by ASTIA between April and July 1962. There were 306 ASTIA descriptors and 87 NASA Unite rms (subject headings and subhead- ings broken "into single units usually consisting of one word which can be coordinated either man- ually or by machine in a variety of different combinations . . ."). 348 Of the terms assigned, 52 were equivalent within each system, giving equiv- alencies of 17 and 60 percent respectively for the ASTIA and NASA terms. 3111 Ibid, p. 76. 347 Painter, 1963 [445], p. 79. 318 Ibid., p. 74. 81 The ASTIA rate of equivalency ranged from 11 to 19 percent when compared with AEC head- ings in the duplication check by Painter and the supplementary Datatrol study of Jaster, but in the basic Datatrol study the ASTIA rate jumped to 28 percent when the entire list of ASTIA de- scriptors was used, including obsolete terminology. The AEC rate of equivalency, however, ranged from 34 to 72 percent as compared with the 15 per- cent indicated by the basic Datatrol study, which "is figured on a theoretical basis of the total term list only part of which is in current use." 348a Painter's conclusions are that although the figures are too sparse for firm convictions, (1) most of the systems can be converted into the ASTIA descriptor structure but not into any of the others ; (2) since equivalency of indexing terms between ASTIA and AEC was not more than 28 percent of the larger system (ASTIA) but was as high as 72 percent of the smaller (AEC) , the smaller could be adapted to the larger but not vice versa; and (3) "with only 60 to 70 percent consistency of indexing within each system and equivalency of only 30 percent within the broadest system, a table of equivalents is at present of little value in either a manual or a machine system." 349 She contends that in order to apply a table of equiva- lents, both a high degree of indexer consistency and a high degree of equivalency of terms used are essential. The Information and Documentation Center (CID) of EUEATOM, an international organiza- tion located in Brussels, Belgium, has published its work of transforming the Subject Headings of the U.S. Atomic Energy Commission (TID-50001, third edition) into a thesaurus of key words. The thesaurus is used for the storage and retrieval of information regarding physics, reactor technology, protection from radiation, application of isotopes, materials, electronics, radiochemistry, and radio- biology. CID reduced the 1960 AEC subject heading list of 12,809 terms to about 4,500 key words, "known also as descriptors, selectors or subject headings . . . ." 350 This reduction was accomplished in large part because of the methods used in the AEC list for indicating alloys and phase studies of mixtures and also because only 7,208 terms were used in the 1960 cumulative index for Nuclear Science Abstracts. CID split up the 4,500 key words into about 50 groups, according to subject field, in order to collect the synonyms and the generic and subordi- nate terms and examine their contents compara- tively. Graphic displays were made for the groups of key words which permitted grouping terms of like association together. "The related terms were linked by lines or arrows, and this provided an impressive and lucid representation of the 'see- also references' included in the American subject headings list." 351 Eventually the organic com- pounds (1,854 key words) and isotopes (1,386 key words) were excluded from the list, leaving a residue of only 1,127 key words. Reference has been made to the desire of the Committee on Scientific Information for an over- all index to the report literature of the agencies of the U.S. Government. Science, Government and Information, the January 1963 report of the President's Science Advisory Committee (the Weinberg report) , finds with respect to a system for handling scientific or technical information in a network of separate subsystems, that "rapid and efficient switching between the different ele- ments of the system is essential," 352 that the first among the means for rapid switching is compati- bility and that, at least so far as title formats and key words are concerned, to be compatible means to be "easily intertranslated." 353 The National Library of Medicine is an ex- ample of a governmental agency planning for both centralization and decentralization. The NLM hopes that its MEDLARS medical bibliographic system will meet a rise expected in demands for service, especially demand-search requests. The Library sees two ways to meet the increased load : First, "increase the capacity of MEDLARS to process the requests, coupled with an increase in communication lines, and possibly coupled with remote input/output facilities located at centers of large demand. Second, decentralize the serv- icing of requests by using remotely located search facilities: the acquisition, indexing, conversion, and storage for retrieval would be done at NLM while magnetic tapes comprising the files would be reproduced by the Library and distributed to remotely located search centers." 354 There is no mention here of centralization of NLM's informa- tion with that of any other governmental agency. Rothgeb (1963 [496]) reports that the Scientific and Technical Information Facility operated by Documentation, Inc., for NASA, "is designed to maintain a centralized processing system for max- imum decentralized use." Rothgeb claims that the centralized processing (1) enables NASA field centers to contribute documents to the system for ready availability to other users of the system, (2) provides the field centers with their own search tapes, and (3) permits searching capabilities which cover information from journals before they are printed and before the search tapes reach the field centers' search system. There is no mention of centralization of NASA's information with that of any other governmental agency. A report to the National Science Foundation by Arthur D. Little, Inc., on centralization and documentation recommends that before elaborate 3480 Painter, 1963 [445], p. 82. 349 Ibid, p. ix. a™ Rolling, 1962 [491], p. 2 351 Rolling, 1962 [491], p. 3. 352 President's Science Advisory Committee, 1963 [464], 353 Ibid, p. 36. ^National Library of Medicine, 1963, [412], p. 61-62. p. 4. 82 word thesauri are developed for existing partially centralized information retrieval systems, there should be more investigation of the statistical tech- niques for generating thesaurus lists automatically and for automating functions performed by hu- man intermediaries. 355 Existing large-scale co- ordinate retrieval systems are regarded as scaled- up and computerized versions of punched card searching systems based on index term matching ; their overall performance in terms of precision and recall measures appears low and inadequate to provide a thorough searching facility. 356 The Arthur D. Little report was made at the behest of the Foundation for the Panel on Science Infor- mation of the President's Science Advisory Com- mittee, as reported by Weinberg. 357 We should note here that Warheit in his review of the Arthur D. Little study (1962 [629]) be- lieves that "the essential conclusion that a large general collection cannot be searched with any real precision or completeness just doesn't stand up. It can be shown theoretically, it is said, that the models developed (in the study) show the inability of making a precise and complete bib- liographic search in a file of over 100,000 items. Yet librarians almost every day make effective searches in much larger files." There appear to be logical questions, too, as to the ability to handle all scientific disciplines in the same way. For example, the Summary of Discussion on Area 7 at the 1958 International Conference on Scientific Information reported by Clapp and Rogers (1959 [121]) reemphasizes the points made by Boutry of the Abstracting Board of the International Council of Scientific Unions theorizing that the many problems in the docu- mentation of the biological disciplines may be ac- counted for by their magnitude, amounting to about 50 percent of all abstracted literature, by contrast with one fifth to one sixth for physics, and less than one third for chemistry. This is cited in connection with the thought of merging the vocabularies for the physical and life sciences. Hyslop contends, on the other hand, that there are similarities in the system of the American So- ciety for Metals at Western Eeserve University, designed for computer processing and a central- ized information service, and the system of the AIChE, designed primarily for manual or simple card files and for use by individuals, but adapted for machine use as well. She believes that "the similarities between the two systems lie in philoso- phy or methodology used to analyze the informa- tion contained in the original document — whether this analysis is done manually or by machine. Both systems utilize three devices for subject analysis. In the ASM-WRU system these three devices are designed as 'code dictionary,' 'role indicators,' and 'punctuation.' In the AIChE system the corresponding terms are 'thesaurus,' 'role indicators,' and 'links.' " 358 Brenner and his associates of the American Pe- troleum Institute report that the centralized in- formation retrieval system for the petroleum industry makes use of a thesaurus, its Subject Authority List, which is "similar to those of the AIChE, EJC, and DDC." Further, ^'no other one vocabulary contains all the technical terms used in the publications covered by API Abstracts.'''' 359 Several representatives of the Federal Govern- ment have proposed a centralized system of code cataloging for machine retrieval. Langenbeck of the Naval Ordnance Laboratory suggests that "a DOD-wide standard code and dictionary can be realized by comparing individual library codes of activities using machine retrieval and arriving at a compromise. A very carefully drawn stand- ard and a control system will be necessary for periodic revisions of the standard code and dic- tionary." He proposes also for "each activity to do the descriptive cataloging, subject cataloging, and machine coding for all reports it creates and distributes. This information is to be incorpo- rated as an integral part of the report on a spe- cific page reserved for library use." 360 Hooker hopes that experimentation on the bib- liographic control of scientific literature "will lead to some standardized or centralized systems which will put all this information on magnetic tape and that we, the users, can buy the tapes and each do his own retrieving. . . . the logical solution seems to me to be preparation of the product cen- trally with the reference use or retrieval of the information locally." 361 Three other citations, already made, may be re- ferred to again regarding difficulty in centralizing indexing vocabularies. Ball (1947 [44]) favors a system for classification, such as that of the Patent Office, integrated with the UDC and he would have a corps of specialists supervising this system continually. Pings (1960 [458]), on the other hand, would have nothing to do with at- temps at a universal classification, contending that there will never be agreement upon a univer- sal classification unless there is agreement in deal- ing satisfactorily with limitations of scientific inquiry. The UDC, too, has had its detractors such as Wellisch (1960 [634]) who finds that many users of the UDC do not know of the changes made in the classification and do not have interest to abide by changes made by the many committees and contributors. Such varying opin- ions certainly have bearing on the success of at- tempts to interchange or centralize indexing vocabularies. 866 A. D. Little, 1963 [344], p. 4. 350 Ibid, p. 2. "" Weinberg, 1962 [636], p. 4. 358 Hyslop, 1963 [278J, p. 293. ^Brenner, et al., 1963 [78], p. 290-291. 300 r.inKenbeck, 1962 [328], p. 13, 17. 3 °» Hooker, 1962 [271], p. 17. 83 3.8. Storage, Search, and Selective Dissemination The scope of this report includes storage in some aspects — standardization of formats, com- patibility and convertibility between forms of storage media, maintenance of cooperative depos- its, and search, in the sense of what is obtained as product of a search. The need for considera- tions of convertibility or compatibility is espe- cially pressing for mechanized information sys- tems. Conventional manual systems require con- sideration of cooperative and coordinated action but in the realm of analysis, identification, and systematization of indexing processes, as we have already seen. Mechanized systems must also con- sider coordinated action in the matter of type and form of storage medium and of search product. Current operations in this area are not extensive. Some proposed activities have been formulated, but we can review the considerations and needs for cooperative steps along these lines. Hayes points out that not only does the mech- anization of information handling processes require "a detailed specification of every minute step and operation" but also that the costs of mechanization are great enough so that the pro- cedures which must be specified in detail must also be highly efficient." 361a This means, among other things, that the inefficiencies of reediting, retrans- scription, reformatting, and the like should be avoided when materials processed in one organi- zation are also to be processed by another. Wherever cooperation, exchange, and interchange are involved between two or more organizations, at least one of which uses a machine system, effec- tive cooperation will thus require the development of appropriate measures of compatibility, as well as the making available of minutely detailed specifications. If, for example, either DDC or the Clearinghouse for Federal Scientific and Techni- cal Information wishes to make copies of mag- netic tape files or of paper tapes used to print announcement bulletins available to other organi- zations, the copies cannot be used elsewhere un- less they are accompanied by specifications in such fine detail as whether descriptors in a list of de- scriptors are separated by two or by three blank spaces and the different sort- sequences resulting from interpretation in different machines of the bit patterns used to encode characters. We must then consider questions of convertibil- ity from one machine-usable format and/or char- acter set to another, from one machine language to another, from a single problem-oriented lan- guage to more than one machine language or in- terpretative program, and from one notation sys- tem to another. In addition, there must be con- sidered the question of automatic conversion from one indexing language to another, the question of automatic conversion from natural language to a documentary language, and the question of auto- matic conversion from natural language to ampli- fied searching languages. Let us turn our attention to two forms of stor- age which involve cooperative action by informa- tion systems and centers. The first form is the cooperative depository library, wherein participat- ing libraries and centers store their cooperatively acquired collections, their little-used material, or even items they no longer need but wish to make available to other organizations. We have already mentioned depository libraries established and maintained by a central facility, such as the AEC network. These also constitute cooperative stor- age. Such cooperative storage facilities could alle- viate many of the weeding-out problems by pro- viding temporary disposals subject to recall as needed. "Weeding requires a great deal of staff time in selecting the items to be discarded, chang- ing the catalog records, actually removing the books from the shelves, deciding how to dispose of the books, and observing federal regulations on the disposal of government property." 362 Kemov- ing obsolete works from a collection may have little significance so far as space saving is con- cerned, but it does expedite research by facilitat- ing access to the more valuable material. The second form of storage we might consider is microphotographic storage — microfilm, micro- fiche, cards and the like. Such media relate mainly to the improvement of physical accessibility; in themselves they do not contribute to the solution of content accessibility. Microphotographic ma- terial can be acquired and stored by an individual library or cooperatively by several libraries; the latter possibility is of greater interest to us here. One point to note in regard to cooperative acqui- sition of microforms is that microfilms made from master negatives do not go out of print, and may therefore be acquired when and if needed. Fussier emphasizes that "mircofilm can economically be acquired either as an edition copy or as a single acquisition to meet an existing need — in contrast to acquisition in anticipation of a future need. The gains in the latter are very great, for it means that large masses of highly specialized and pe- ripheral literature — systematically and coopera- tively collected- — would be available on a deferred basis, when needed, at costs less than multiple acquisition, cataloging, and storage." 363 A cooperative program that combines depository storage and microforms is that for "auxiliary" or "demand" publication in which documents are deposited in a central repository and from which copies can be had on demand. This requires an announcement through published summary or con- densed article of the availability of the deposited document. The technique of auxiliary publication already has been used for distribution of U.S. gov- seia Hayes, 1960 [241], p. 6. 302 Evans, 1963 [188], p. 5. 303 Fussier, 1951 [207], p. 82. 84 ernment research reports by OTS and theses by University Microfilms, Inc., and in the ADI's "auxiliary publication program" 3 wherein tabular data and extensive texts not ordinarily published by scientific journals, but useful for a limited num- ber of specialists in the field, are placed on micro- film. They, along with microfilms of certain rare and out-of-print journals, are transferred to the LC Photoduplication Service where photocopies are made available upon request. University Microfilms, Ann Arbor, Michigan, maintains files of many publications in microfilm form, and will duplicate its files in either micro- film or hard copy. There are five categories which reflect the major interests of this organization: periodicals, newspapers, books, special fields, and equipment. The periodicals include microfilms of modern periodicals, English literary periodicals of the 17th, 18th, and 19th centuries, Russian- language periodicals, American periodicals of the 18th and 19th centuries, and Chinese-language periodicals not previously available in the United States. Newspapers on microfilm include Irish newspapers prior to 1750 in Dublin libraries, se- lected English newspapers, modern newspapers (including the New York Times, Osservatore Romano, and Commercial and Financial Chron- icle), and Arab newspapers. Books on microfilm include English books of the 15th, 16th, and 17th centuries, scarce manuscripts and books on Amer- icana, and out-of-print books including Russian material. Among the efforts to control microfilm material bibliographically have been "the Union List of Microfilms, published by the Philadelphia Biblio- graphical Center, the Microfilm Clearing House maintained at the Library of Congress . . . ." 364 Also, "The Navy Research Section's experience with the Microcards has been in connection with an experimental program proposed late in 1950 by the Office of Naval Research . . . distributing ... a particular type of Microcard for the Un- classified and Restricted reports abstracted in NRS' abstract bulletin — the Technical Informa- tion Pilot (TIP). This card consists of a con- ventional Microcard laminated to the back of a standard NRS catalog card." 365 All the problems associated with the use of mi- croforms have not been solved. As Fussier points out, ". . . In many respects photographic access is of critical importance in relation to the success of any broad plan for the cooperative acquisition of original materials . . . [but technical advance- ments will be required to] increase accessibility by promoting ease of use, by reducing costs, and by faster availability." 366 Three government agencies — the AEC, NASA, and DOD — have reecntly standardized a micro- fiche system for reproducing all research and de- velopment reports from these agencies and their contractors. The microfiche is a negative card containing document pages in micro form. The agencies have not only agreed to the outside di- mensions of approximately 4x6 inches, but to other standard dimensions as well — spacing be- tween pages, reduction ratio, and others. These standards make it possible for documents from the three agencies to be viewed and reproduced by the same equipment, viewing screens and projection lenses. Dissemination of reports will be faster and simpler, and the agencies will be able to cut costs. The Federal Council for Science and Technol- ogy adopted the standard in the Spring of 1963 for government- wide use. 307 There is also the pos- sibility that central Federal support should be provided for the procurement (or development as necessary) of inexpensive readers capable of han- dling the new standard microform and their place- ment in all Federal libraries and in all cooperat- ing public or private libraries. Assuming that storage, subject to selective re- call and retrieval, has been effectively achieved, what is the situation with respect to the compati- bility or convertibility of the search product? The product that is obtained depends largely on the type of system from which it comes. For ex- ample, some systems store data, technical or ad- ministrative in nature, and process this material to produce factual results in the course of normal operations or in answer to specific queries. Chem- ical compounds are tested for biological activity and the results recorded in a data-storing system ; then studies are made of correlations between chemical structure and biological effects. The products of such a system are actual facts and figures on correlations, structures, and activities. On the other hand, some systems store refer- ences to data, or to the documents containing data, and produce notations of pertinent references in answer to particular queries. The number of such systems is large and their operations are highly diversified with reference to the degree of mech- anization employed in their operations. Still other systems store data or references but produce general search aids as part of an integrated oper- ation. Search aids include indexes, biblographies, and publication aids providing, for a specific area of science and technology, a means for individuals to make particular searches or to maintain cur- rent awareness. Alternatively, the product of an information system might be a machine-searchable file, such as the reproduction of decks of punched cards, with which the individual can set up his own system for making specific searches. In all cases, products of information systems, whatever their form, need to be considered in terms of their compatibility or convertibility one to another. The question is whether these prod- 304 Council on Library Resources, 1958 [140], p. 19. 3M Grav, 1952 [2231, p. 58. ™ Fussier, 1951 [207], p. 183. 3I » "Microfiche Standards Adopted," 1964 [3S3]. 85 ucts can be used by others than those who pro- duced them, and whether they can be exchanged and incorporated into other hies. The question, in general, has been explored elsewhere in this re- port. Here, we shall mention only those few fur- ther illustrations of possibilities of compatibil- ity in the area of the production of selective bibliographies. Betty Gray of the Textile Fibers Dept. of E. I. du Pont de Nemours uses 36S the tape typewriter in the preparation of annotated bibliographies to reports. A punched paper tape, produced when reference cards are typed on an automatic type- writer, is used for type-out of a final master copy of the bibliography. The sections of the punched tape corresponding to individual references can be arranged conveniently to satisfy the final arrange- ment desired, and the searcher is not confined to a predetermined order in making the search. The use of this system results in a saving of 60 percent over the time required for manual typing of a bibliographic report. A further comment may also be of interest : "Not only is the ability of federal librarians to prepare useful bibliographies most inadequately utilized, but those which have been prepared have generally been issued in quantities too small to reach all who would benefit from them. A coordinated biblio- graphic service, established through the coopera- tive efforts of departmental libraries, would do much to improve this situation. Such a service should, of course, complement and assist rather than duplicate the bibliographic work of nongov- ernmental bodies so as to promote broad, high qual- ity, national bibliographical coverage." 369 Selective bibliographies are related on the one hand to specialized mailing lists and on the other to the selective dissemination systems first sug- gested by the Luhn (1958 [355], 1962 [353] ) . An early Government example of selective mailing lists was that of "the Solid Propellant Informa- tion Agency (SPIA), which was established in 1946 at the Johns Hopkins University Applied Physics Laboratory ... A unique feature of SPIA's operation is its Joint Army-Navy- Air Force Mailing List for the distribution of solid propellant technical information." 370 Another co- operative defense agency example was the Army— Navy-Air Force Guided Missile Mailing List which had "the effect of grouping of individuals working on similar research and providing them with information of mutual subject interest." 371 Selective dissemination, or SDI, systems are tailored instead to individual subject interests, particularly the research worker's current aware- ness needs, where he has filed a "profile" of his current subject interests which can be updated by machine as a result of feedback from him. Incom- ing items that "match" his stored profile result in SM Gray, 1961 [220], p. 71. 369 Evans, 1963 [188], p. 30. 3 '°Herner and Herner, 1959 [254], p. 190. 3n Heald, 1952 [244], p. 140. notifications being sent to him of the availability of these items on a turn-around card so that he may order any of the items. Kraft, reporting in 1963 on the then state-of-the-art of SDI systems makes the following comments : "Those people who have a current awareness need should furnish lists of words which charac- terize their various interests. These are the same 'hot-button' words which the reader looks for when he browses . . . "At the time of this writing, there are in excess of a dozen SDI systems in the country . . . "Efficient dissemination, when a document first appears, reduces the retrieval problem later. SDI is a kind of 'retrieval-in-advance' — a current awareness program that helps users keep abreast of new information . . . "An analysis of users' replies indicated that two- thirds of the notices were 'of interest' and one- third of 'no interest.' Considering that a 'typical' user received an average of five SDI notices per day, the noise level of 34 percent was tolerable, since it required less than one minute to read an SDI abstract and punch out the reply card." 372 At NASA, an SDI system is under development and experimental trial. "Reports having a prob- ability of specific interest to a particular individ- ual are selected by specially prepared computer programs. Each man's 'interest profile,' consisting of subject terms and phrases related to his activi- ties and interests is compared with the subject in- dexes of all reports to be annotated in current is- sues of STAR. When an optimum match between terms occurs, an abstract of the selected report is mailed directly to the participant, delivery being made even before the abstracts are available in STAR." 373 3.9. Standardization In spite of our earlier statements to the effect that too little progress has been accomplished and too little effort is being expended on problems of cooperation, compatibility, and convertibility, we can report considerable efforts toward agreement and standardization in progress. Such efforts are being carried forward by or- ganizations primarily devoted to standardization, by societies devoted to professional activities but vitally concerned with standards, and by other groups whose work is affected by and would benefit from the development of standards. These stand- ards are concerned with hardware, equipment, physical formats, and the like ; they are also con- cerned with intellectual aspects, languages, codes, subject contents, and similar points. Many of these have been touched upon in earlier sections of this report. Here we shall discuss the standards and standardization efforts in terms of the groups working on them. 3 ™ Kraft, 1964 [323]i, p. 143-151. 373 Scientific Information Notes 6. 1, 1-2 (1964). 86 3.9.1. Programs of Federal Government Agencies Activities of the Group for the Standardization of Information Services were mentioned earlier in this review : four government agencies ( AEC, NACA, Central Air Documents Office, and LC Navy Kesearch Section) came to an agreement regarding a few of the problems encountered in their bibliographic services. The problems were attacked in meetings of working-level individuals by first setting up standardization goals, the at- tainment of which would permit each agency to use the products of the others. Following dis- cussion, the tentative decisions reached were taken back to the appropriate authorities in each agency for approval, which was generally quickly forth- coming. The major items on which agreement was reached included the common format for li- brary cards and the development of a subject heading list that would be as nearly common to all agencies as was possible. In 1956 Connor, of the Technical Library of the U.S. Naval Air Missile Test Center at Point Mugu, Calif., in citing the need for standardization in the documentation of government research reports, asked specifically for a "statement setting forth physical and bibliographic requisites in the prepa- ration and publication of reports." 374 This was to be one of the specifications within a government contract. The items which Connor wished to see standardized were (1) with reference to physical format, the size and kind of paper on which the document is inscribed, and the types of cover and of binding (e.g., whether stapled or sewed) ; and (2) with regard to bibliographic data, title pages were to show only company or agency along with personal authors, dates including inclusive dates for periodic reports, title, report number, and address of organization responsible for the report. The attention of military agencies to standardi- zation and compatibility is demonstrated by the activities of the DDC which include work on ab- stract standardization. "A uniform abstract cov- ering all bibliographic elements for technical re- ports is being devised for use by originators and authors, thus paving the way for automation on common grounds and large savings in time, ex- pense, and repetitive bibliographic efforts." 375 Col. Vann, then director at DDC (ASTIA) , spoke to this point: "[It is] important ... to start at the point of generation with a format and stand- ards which will speed up the cataloging, indexing, and retrieval of a document when it reaches the Documentation Center." 376 Another area of DDC effort is concerned with interchange of magnetic tapes. The prerequisites here are an essentially complete automatic data processing system and uniform processing of in- formation. Exchange of compatible tapes will then be feasible and permit rapid interchange of information with minimal duplication of effort. 3 ™ Connor, 1950 [134], p. 155. m ASTIA statement in NSF-G3-5 1962 [418], p. 66^69. 8" Vann, 1963 [606], p. 221. Heald emphasizes these points (1963 [175]) when he states that ASTIA has made efforts toward standardization and compatibility in the revision of its Thesaurus of ASTIA Descriptors and in the desire for adoption of abstract stand- ardization. He lists the entire bibliographic cita- tion for describing a report as being one of the elements of a standard abstract, along with the summary or body of the abstract. This standard abstract includes title, author, originating agency, identifying number, and subject indexing, as well as the summary. Heald contends that by means of coordination within the abstract standard there will be attained (1) compatibility of indexing and abstracting; (2) computer-to-computer communication through interchange of magnetic tapes and rapid network transmission; (3) reduction in the time for proc- essing information, between its origin and its being made available; and (4) reduction of the cost of index and abstract processing: ". . . tremendous gains are possible by building a vocabulary as compatible as possible with other information ac- tivities, such as those of NASA, AEC, OTS, and DOD activities and contractors. The eventual pay-off here is computer-to- computer interchange of information and interchange of magnetic tape." 377 Similar opinion was expressed by Hall, who said, "Standardization in such matters as report designation, layout of covers and preliminary pages, the furnishing of abstracts and catalog cards within the reports themselves, bibliographic usage, arrangement of contents and indexing ac- complishes two important and money-saving ends. It provides for a more rapid and accurate production of the reports themselves . . . and at the other end of the cycle provides for faster and more effective indexing, cataloging, coding." 378 The survey of scientific and technical informa- tion retrieval schemes within the Department of the Army, by Weik and Confer (1962 [633] ) , has remarks pertinent to compatibility and standard- ization. They find that there is very little uni- formity or standardization among such schemes, and that the great tendency to devise codes for easing the classification, storage, and retrieval problem is evidence of the great ambiguity in the English language. The Institute for Defense Analysis made a digi- tal computer applications study for the Advanced Kesearch Projects Agency of the DOD. This study also found that there is lack of coordination in efforts at automation, resulting in functional and technical system incompatibilities and lack of standards for machine languages and program- ming languages, an important factor in converti- bility problems. 379 3 " Heald, 1963 [242],, p. 42. 37S Hall, 1960 [233], p. 62. 370 See "Where the Computer Fits in Command and Control," 1962 [643]. The hearing before the House Subcommittee on Census and Government Statistics of the Com- mittee on Post Office and Civil Service which was held on June 11, 1963, on the use of electronic data processing equipment has several statements rele- vant to standardization and compatibility. The Comptroller General of the United States, Mr. Joseph Campbell, found that (1) lack of compati- bility between magnetic tape systems used in equipment of different manufacturers results, un- less 'special additional equipment is designed to overcome this shortcoming or unless the data on the magnetic tapes are converted to punched card or paper tape media and then reconverted to mag- netic tape for use in follow-on operations;" (2) "costs are incurred because of the need to convert and reconvert files and records from one type of data processing equipment to other models of equipment;" and (3) a central management office is needed to properly exploit the possibilities of Government-wide integration of systems and to plan for their development and growth." 38 ° On August 1, 1963, the same subcommittee sub- mitted to the parent Committee its Interim, Report on the Use of Electronic Data Processing Equip- ment in the Federal Agencies (1963 [584] ). The subcommittee states, "In fact, in almost every prob- lem area identified during the subcommittee hear- ings, contradictory testimony was given. Some agencies, for instance, wanted more centralized control, others wanted less. Some agencies testi- fied that common machine languages (COBOL, FORTRAN, etc.) were timesavers; others indi- cated that they had not, as yet, found these lan- guages particularly helpful. . . . On the subject of standardization and compatibility, some wit- nesses testified that the Federal Government was making substantial progress in achieving greater compatibility, whereas other witnesses believed that considerably greater Federal initiative and effort were needed in this area." These remarks are applicable not only to busi- ness or management data processing systems, but also to the scientific and technical information processing systems with which we are concerned here. The basic form of the bibiographic entry has received the attention of the Committee on Scien- tific Information, as was related in an earlier sec- tion of this report. As a result, a Standard for Descriptive Cataloging of Technical Reports was published in December 1963 by COSI ([195]). There is still the necessity, however, for coding the items of the entry in order to make possible the "ordering out" or retrieval on those items. We have also mentioned earlier the efforts of Federal agencies to standardize the form and use of the microfiche. Thus an agreement was reached in 1963 by NASA and the AEC whereby a uniform ratio of reduction in microphotocopying would allow the use of the same reading and printing equipment on all microfiche versions of technical reports regardless of the agency origin of the re- ports. This agreement was extended to OTS (Clearinghouse) for use in its large program for distributing DDC reports as well as other govern- ment-generated or -acquired material. Another event of interest in the field of stand- ardizing microforms is the publication of stand- ards and specifications" for microphotography of DOD engineering data. The DOD has issued throughout the years such standards and specifica- tions, the most pertinent of which were assembled and published in 1963 by the National Microfilm Association. These documents, according to the Association, "supply basic data as it relates to the microphotographic reproduction of engineering drawings and related materials in accordance with the requirements of the Department of Defense." 381 In 1962 the DOD was given the annual award of the National Microfilm Association as its ma- jor recognition for outstanding service in the field of microreproduction. In the citation for the award the DOD is credited with developing and introducing "new and improved concepts of re- cording and reproduction of engineering docu- mentation ; the collection, interpretation, coordina- tion, development and publication of standard specifications; industrywide acceptance of stand- ardization of procedures and nomenclature. . . ." 3.9.2. Programs of the American Standards Association (ASA) The American Standards Association is a pri- vately supported organization "acting as the na- tional clearinghouse and coordinating agency for voluntary standards in the United States." 382 Its main functions are to provide systematic means for development of standards, to promote their use, to coordinate standardization activities in the United States, and to act as a clearinghouse for information on standards. "An American Stand- ard is a voluntary national standard, arrived at by common consent, and available for voluntary use." 383 ASA Sectional Committee X3 on Computers and Information Processing has seven working subcommittees. Subcommittee X3.1 is on optical character recognition (OCR) , having task groups on font development, printing, and applications. Subcommittee X3.2 is concerned with coded char- acter sets and data formats and has one completed standard, American Standard Code for Informa- tion Interchange (1963 [23]). This American Standard, 3.4-1963, was approved by the ASA on June 17, 1963, with the Business Equipment Manu- facturers' Association as sponsor. The standard presents the standard coded character set to be U.S. Congress, 1963 [5S5], p. 7-9. 381 National Microfilm Association, 1963 [414], p. iil. 382 Henriques, 1964 [251], p. 23. s^Utman, 1963 [605]., p. 9. 88 2. STANDARD CODE . °0 1 1 i i 1 °0 1 1 'o 1 1 1 ^ s b 4 b 3 * >>2 ( 1 2 3 4 5 6 7 NUL OLE SP v p P 1 1 SOH 0C1 ! 1 A q 1 2 STX DC2 " 2 B R b ' 1 1 3 ETX 0C3 # 3 C s c s 1 4 EOT 0C4 $ A D T d t 1 1 5 ENO NAK % 5 E U e u 1 1 6 ACK SYN a 6 F V f V 1 1 1 7 8EL ETB ' 7 G w 9 w 8 as CAN ( 8 H X h ' 1 9 HT EM ) 9 I Y i y 1 10 LF SS # : J z i 2 1 1 1 1 VT ESC + 5 K c k i i 12 FF FS , < L "" I ' i 1 13 CR GS - = M ] m ) i 1 14 SO RS . > N ~ n 1 i 1 1 15 SI US / ? _ DEL used for information interchange among informa- tion processing systems, communication systems, and associated equipment. The standard specifies the correspondence of seven-bit patterns with 128 graphics and control codes; the graphics are for printable characters, and the controls are to initiate appropriate equip- ment functions. The 128 -character code set cov- ers 36 controls, 28 special characters, 10 numerics, and the 26 alphabetic characters, leaving 28 bit patterns unassigned. For example, the 7-bit posi- tional order for the letter "B," from the high- to the low-order position is 1010010. The standard does not define the means by which the coded set is to be recorded in any physical me- dium nor does it include any redundancies. It does not define techniques for error control. It does not specify a standard collating sequence. The code is referred to in the trade as "ASCII," pronounced "asky." (See figure 2.) Bemer, in his first article on ASCII for Data- mation, states that it may be expanded to 8 bits and that representations in various media are not defined. "The set can be collapsed in a regularized and prescribed manner, if required, into a 6-bit set for existing 6-bit machines and other equipments, to a 5-bit set for modification of existing Teletype and Telex sets (particularly in Europe), and even to a 4-bit set." 384 In his second ASCII article Be- mer states that the code demands that "whenever the computer talks with strange equipment, not of its own kind, that it do so through the medium of this code." 385 Hence, each machine will have to talk only with ASCII instead of the other inter- nal codes, numbering now about sixty. Bemer lists the following advantages for com- puter programming purposes: (1) manipulation of graphics by classes; (2) fewer instructions in scans because of regularity and unique codes; (3) faster and cheaper sorting when the collating sequence is identical to the binary sequence of the codes for the graphics; (4) reduction in the num- ber of routines required to be programmed, par- ticularly for satellite equipment ; (5) fewer tables for mixed codes in communications, particularly those controlled by store-and-forward message switching systems; (6) clarity of printed output, particularly the reproduction of the source pro- gram in the printed record of processing; and (7) a tendency for keyboards to be identical with typ- equip Ion «f ment. mg communications Bemer comments also ' r that some other internal code phis the translation mechanisms required might be more economical for some equipment then would ASCII internally. . . . The new code has so many inherent economies that it might pay for the redesign itself. IBM has perhaps the least problem of any manufacturer; with 9 differ- ent codes already in their various computers, ASCII presents only 11 percent additional prob- lems." 386 The International Business Machines Corps, in the August 7, 1963, special issue on ASCII of its IBM Data Processor states that "IBM plans to provide whatever means are practical to meet cus- tomer needs for using the standard code as soon as possible after media standardization is ap- proved by the American Standards Association." To inject a less serious tone, perhaps, we note that "the Federal Government, the military serv- ices, the manufacturers' society, and the computer community are now engaged in an extended de- bate on just what the characters in a 6-bit set should be and how these characters should be en- coded within the set. There are already almost more variations on this basic set than there are rabbits in the Western United States." 387 Subcommittee X3.2 of the ASA Sectional Com- mittee X3 is also actively engaged in concern with the problems of input-output media, including a code for perforated tape and specifications for the physical dimensions of both 1-inch and H/ieth-inch perforated tape; codes for magnetic tape and punched cards; codes for both perforated and magnetic tape for numerical machine tool control ; and. with specifications for the physical dimensions of i/o-inch magnetic tape and for the 80- and 90- column punched card and the edge-punched card. Subcommittee X3.3 is concerned with digital data transmission. It proposed the first American standard in information processing, naling speeds for data transmission that on sig- . Standard X3.1-1962 was approved August 1962 by the ASA (1962 [26]) with the Business Equipment Man- ufacturers' Association as sponsor. The estab- lishment of the speeds was considered necessary to ensure compatibility between communication channels and data terminal equipments of com- munications systems. The standard signaling speeds are 600, 1200, 1800, 2000, 2400, and 3000 bauds. The statement concerning signal element duration is that "a synchronous signal train at the 381 Bemer, l!)f>3 [49], p. 35. 385 Ibid, p. 40-41. k» Ibid, p. 40. 387 Patrick and Black, 1964 [450], p. 31. 89 interface between the data communications equip- ment and the data processing terminal equipment, after synchronization is established, shall consist of a sequence of marking and spacing signals, all of which are integral multiples of the signal ele- ment. The signal element duration is equal to the reciprocal of the signaling rate in bauds." Subcommittee X3.3 has six task groups on liai- son, glossary, data transmission formats, end-to- end control characteristics, system performance and digital data transmission speeds. Subcom- mittee X3.4 is concerned with common program- ming languages, to be discussed in section 3.9.4. Subcommittee X3.5 is on terminology and a glos- sary for information processing and hopes to pro- duce a proposed American standard glossary soon. Subcommittee X3.6 is on automatic data proc- essing problem description and analysis and has a proposed standard which prescribes and defines the symbols used on flowcharts to represent both the sequence of operations and the flow of data and paperwork of information processing systems. The basic symbols include those for input/output, processing, flow direction, and annotation. Spe- cialized input/output symobls are those for the punched card, magnetic and punched tapes, docu- ment, manual input, display, communication link, and on-line and off-line storages. Specialized processing symbols include decision, predefined process, manual operation, and auxiliary opera- tion. There are additional symbols for a connec- tor and a terminal. (These symbols are somewhat reminiscent of the graphical drawing symbols allowed by the U.S. Patent Office, in its Guide for Patent Draftsmen, for use on patent appli- cations. The symbols recommended in the Guide are for common materials, such as metal or cork, electrical elements, and mechanical elements). Subcommittee X3.7 is on magnetic ink char- acter regonition (MICE) which is closely related to the optical character recognition interests of X3.1, taking over earlier work by the Office Equipment Manufacturers Committee in conjunc- tion with the American Bankers Association. It is interesting to note that Sectional Commit- tee X3 (SC X3) of the ASA was organized in 1960 to work on standards for software (program instructions and operating procedures for infor- mation processing equipment) more than 18 months before Sectional Committee X6 (SC X6) was organized to handle standards for hardware (equipment, especially information processing equipment) . Sectional Committee Z39 of the ASA covers Library Work and Documentation. Its purpose is to develop standards for concepts, definitions, terminology, letters and signs, practices, methods, supplies, and equipment used in the field of li- brary work and the preparation and utilization of documents. The committee is sponsored by the Council of National Library Associations. Sec- tional Committee Z39 in the first quarter of 1963 was giving particular attention to four areas : ab- breviations for periodical titles, abstracts, trans- literation, and trade catalogs. The standard pro- posed for the transliteration of modern Russian letters had been developed by the Subcommittee on Transliteration in close cooperation with the Royal Society of London. The Subcommittee on Abstracts of the ASA SC Z39 regards an abstract as an abbreviated rendition of the significant content of a scientific paper or report, book or other documentary item. The intention of the proposed standard on ab- stracts is to comment briefly on the function of an abstract, its character, content, style, and lan- guage, and on the publication of abstracts by abstracting services. There is also a Subcommittee on Classification, which held its first meeting in November 1963. The Subcommittee on Transliteration of Z39 is concerned with transliteration from one language alphabet to another: Greek to Roman, Russian (Cyrillic) to Roman, Chinese to Roman, etc. Ac- tivities of this committee are aimed at develop- ing a system acceptable to librarians and infor- mation specialists in the United States, with hopes for coordination with similar activities in organizations abroad. The subcommittee aimed, for example, to present a single new system of Cyrillic transliteration for modern Russian only ; having worked on this program for some time, it turned to consultation with British representa- tives on the needs and aims of transliteration pro- grams overseas. The need for standardization in this area is ob- vious, especially in light of the ever-increasing and ever more important foreign language mate- rial to be processed. (The question of language problems as such is discussed in section 4.1.2 of this report.) The transliteration problem touches not only conventional information handling sys- tems but mechanical translation research pro- grams also. Interests of the latter were expressed by Kellogg when he said, "If we [U.S., Russian MT programs] could agree on two simple things, we would both be happy. One is a uniform trans- literation system, just for changing people's names from one alphabet into the other. We have a sys- tem by agreement within the Western World, the Russians use a different one. "Secondly, if we could agree to have them print a little box instead of a period at the end of a sentence, and a comma that is a little more definite than the kind we have now, and if we traded them journal for journal, this would save us trouble — but we can't get agreement." 388 However, at- tempts to reach such agreement do continue. Some standards have been promulgated suc- cessfully. Thus, American Standard Z39.1 was sponsored by the ALA, approved on June 7, 1943, and reaffirmed in 1959. It concerns the manner by which periodicals are issued, numbered, and 338 U.S. House, 1960 [588], p. 61. 90 named, and the manner by which their pages are numbered ; it requests that each periodical include a title page, table of contents, and index. American Standard Z39.4 was sponsored by the Council of National Library Associations and ap- proved on May 5, 1959. The criteria are for in- dexing books, periodicals, and other documentary materials. An index within the fields of library practice and documentation is denned as "a guide to the contents of any reading matter or other documentary materials which provides a syste- matic, sustained subject analysis of the contents of such materials arranged according to alpha- betical, chronological, numerical, or other chosen order. Each entry is followed by page number, paragraph number, or other indicator showing the exact location of the reference." 389 The word "subject" as used in the above defini- tion is defined as "a unit concept found in, or derived from, the material indexed. The unit concept may be found or expressed as a thematic topic, a name, a date, the first line of a poem, the title of a work, an expression coined to give the gist of the material indexed, etc." 390 The matters covered concerning the content, organization and style of indexes include the entries (headings, sub- headings, and modifications) , the alphabetizing or filing, inversion of headings or names of compa- nies or organizations, references (mainly to pages), typography, and cross references. Z 39.5—1963, "American Standard for Period- ical Title Abbreviations," has also been issued. "General rules are provided in the standard for the omission of letters, punctuation, conflicting title words, and multilingual abbreviations . . . The product of two years' research by a committee of experts, the new standard contains more than 2400 abbreviations for words commonly found in periodical titles." 391 In 1960 the American Standards Association and the International Standards Organization decided to form joint committees on computers and information processing. The groups involved were the ASA SC X3 and the ISO Technical Committee 97 (TC 97). Scopes for these com- mittees "were adopted in organization meetings held in 1960 and 1961. ISO assigned the Secre- tariat of TC97 to the United States and the ASA, which in turn recognized the Business Equipment Manufacturers Association (BEMA) as sponsors of X3 and TC 97, with X3 to develop proposed standards for both the U.S. and the world." 392 Formal standardization work in photography, at least in the United States, was first initiated in 1938 under the procedures of the ASA. The com- mittee established at that time was designated the ASA Sectional Committee on Standardization in the Field of Photography, Z38; the committee 888 American Standards Association, Inc., 1959 [22], p. 5. 300 Ibid. 801 Scientific Information Notes 0, 3, 17 (1964). 8M Utman, 1903 [605], p. 9. operated for more than 10 years under the spon- sorship of the Optical Society of America and was responsible for the development of over a hundred standards. On November 30, 1950, ASA Com- mittee Z38 was disbanded with four new commit- tees replacing it, and a fifth was organized on June 25, 1953. ASA Sectional Committee PH 5 on Photographic Reproduction of Documents was given i responsibility for standards "for photo- graphic materials, apparatus, and processes pertaining to production, use, storage, and pres- ervation of document reproduction." Sectional Committee PH 5 has been of considerable interest m the field of documentation owing to its stand- ards on the various kinds of film. The American Standard PH 5.3 is the one of most direct interest to endeavors m microphotography although it was intended basically to be of interest for motion picture film. 3.9.3. Programs of the International Standards Organization (ISO) Donker Duyvis wrote in 1954, as the Secretary General of the International Federation for Docu- mentation, on Standardization as a Tool of Scien- tific Management. He gives as a positive defini- tion of standardization: "to bring production to a higher level, to guide and plan judiciously the necessary diversity in order to promote harmony m variety, and to assure that human labor will be used in a worthy way." 383 He lists as subjects belonging to standards in librarianship : the ma- terial and layout of documents; the elements and editing of their contents; their arrangement, filino- and storage, and various mechanical devices. Donker Duyvis stresses that ". . . particularly for librarians and documentalists, standardization loses half of its value if it is not done on a world- wide basis." 394 It may also be of interest to note an example for documentation for standards. "The late F. Don- ker Duyvis, former Secretary General of the Inter- national Federation for Documentation, united his interest in providing documentation service for standards with that of the ISO, and became chairman of ISO's Committee for Index Cards for standards that are issued by Member Associations." 395 The International Standards Organization (ISO) had in 1954 a technical committee on docu- mentation called ISO/TC 46. In that year, ISO m collaboration with UNESCO published draft recommendations for the layout of periodicals, for bibliographical citations (the citing of titles of documents), and for bibliographical references (the identification of publications, considered either as bibliographical units in themselves or as parts of larger works) ; it published also ISO/R4, accepted by the ISO Council as its recommended International Code for the Abbreviation of Titles 803 Donker Duyvis, 1954 [176], p. 410. 3M Ibid, p. 426. s^Murra, 1962 [402], p. 407. 91 of Periodicals. These examples are cited as evi- dence of continuing international activity in the standardization of documentation. "Progress in international standardization work in the field of documentation has been steady, if not spectacular ; since the war, there have been five meetings of Technical Committee 46 (Documentation) of the International Organization for Standardization . . . but much of the real work has been done by correspondence and in special meetings of work- ing parties." 396 In 1954 Lloyd ( [346] ) prepared a selective list of standard-practice recommendations available from a baker's dozen of standards-issuing orga- nizations. He noted that of the many problems which plague the technical editor, those which in- volve layout, references, transliterations, abbrevia- tions, proof corrections, and the like can be sim- plified by the adoption of standard practices. The Lloyd list includes seven recommendations for abbreviation of periodical titles, two for abstracts and synopses, six for bibliographic citation prac- tice, five for symbols and abbreviations, and four for Cyrillic alphabet transliteration. The then pending ISO recommendations thought to be most urgently required were: ISO R4 (1954) : Interna- tional code for the abbreviation of titles of pe- riodicals; ISO Draft Eecommendation 23 (1953) : Bibliographical citations; ISO Draft Recom- mendation 24 (1953) : Bibliographical references; ISO Draft Eecommendation 3 (1953) : Layout of periodicals. Lloyd remarked that "some of the most useful recommendations are not as well known as they should be, partly because they are not always readily available and partly through lack of publicity." 397 New "items considered for standards work by ISO/TC 46 at its meeting in Brussels in 1954 in- cluded not only the layout of index cards, proof correction signs, and transliteration of Oriental alphabets, already mentioned, but also two other proposals, which, however, were deferred not only on grounds of overlap with other organizations but also because of serious question as to the fea- sibility of reaching agreement. These were for 'canons' or 'guiding principles' for cataloging and classification rules. With respect to the latter, in particular, the FID General Secretary stated that his organization "had long been studying the prob- lem, and had found it impossible to reach much international agreement yet on such philosophical subjects." 398 In addition, ". . . At the recent Brussels meet- ing of ISO/TC 46, it transpired that criticism of Draft ISO Rec. 24, Bibliographical references (full references) had been so severe as to necessi- tate a complete redrafting . . ." 39 ° More success- ful, though, were "The Code for Abbreviation and •w Lloyd, 1955 [3471, P. 49. 3l "Llovd, 1954 [346], p. 282. 383 Lloyd, 1955 [3471, p. 53. 389 Lloyd, 1955 [347]., p. 50. Layout of periodicals, with Bibliographical cita- tions and Bibliographical references, [which] were regarded by Unesco as sufficiently important to warrant a special grant, and the ISO/TC 46 Secre- tariat has recently discharged its obligation to Unesco by issuing these, together with an intro- duction, in the form of a report to Unesco entitled Standardization in the domain of documenta- tion . . ." 400 In the field of microreproduction, ISO has also been active: "In April 1950 the International Standards Organisation (ISO), meeting jointly with the International Federation for Documenta- tion (FID) at Ascona, established a subcommittee of Technical Committee 46 to draft recommenda- tions for (a) the terminology of microcopies and (b) the size of the microcopies. The meeting of the ISO committee at Paris in June 1955 advanced, but did not conclude, these studies. In 1953 the American Standards Association established a new Committee (PH 5) to develop standards for ma- terials, apparatus and processes connected with the production, use, storage and preservation of photo- graphic documentary reproductions." 401 The ISO has established standards in this area : . . . "ISO Mire {Test object) is made up of a 25 x 50 mm area with lines of artificial 'words' in vary- ing sizes of type, each 'word' consisting of four symbols in the form of a conventional (ISO) octa- gon with two diagonal lines, replacing ordinary letter characters, which can often be guessed at even when not strictly legible . . . ISO Micro- mire (Micro Test Object) is a carefully prepared microcopy (1: 10 reduction) of the ISO Test ob- ject. Strips of the Micro Test object are then mounted in the form of a pattern on a chequered black-and-white ground, and this chart is used for testing image sharpness, etc., in microfilm cameras and reading apparatus." 402 Donker Duyvis wrote (1954 [176]) of his hope for the ASA to cooperate closely with the ISO in formulating standards for reading apparatus. Lloyd reports, "Two new draft proposals referring to the testing of microcopy reading apparatus are now under consideration by the nineteen member countries of the main Technical Committee ISO/TC 46 . . . ." 403 Furthermore, "The subcommittee [ISO/TC 46/SCI Documentary Reproduction Subcommit- tee] unanimously adopted a recommendation from a three-man . . . commission to abandon the terms 'Microcard,' 'Microcarte,' and 'Microprint' as confusing, indefinite and possibly also propri- etary. In place of these, it accepted a definition of the word 'Microfiche' (same word in both French and English . . . ." 404 A few citations to standardization on translit- eration seem appropriate inasmuch as the trans- *>° Ibid. p. 49-50. 401 Born, 1956 [691. p. 168. «2 Lloyd, 1955 [3471, p. 51-52. ■"KLlovd, 1955 T347], p. 51. 401 Ibid, p. 51-52. 92 literation of the languages of the world has always caused so much heated controversy. Frontard, of the French Standards Association, reviewed inter- national efforts toward standardization of trans- literations, beginning with Recommendation ISO/R 9 of the Technical Committee 46 on Docu- mentation of the International Standards Orga- nization ( 1961 [205] ) . ISO/R 9 was published in 1955 and recommended an international system for the transliteration of Cyrillic characters. Progress was reported on the transliteration of Greek and the nonvocalic languages, such as Hebrew and Arabic. At a meeting of ISO/TC 46 in London in June 1960, Chinese, as an ideographic nonsyllabic language, was regarded as incapable of transliteration and amendable only to transcrip- tion, etymologically shifting from writing in one alphabet to writing in another. Neiswender wrote in 1962 that the ISO/R 9 "was a landmark in the history of transliteration, for it represented the first system based on interna- tional agreement, and it secured the approval of 20 linguistically dissimilar countries . . . ." 405 However, the ISO system, although adopted by UNESCO for its bibliographical publications and library catalogs, "did not secure general acceptance among English-speaking countries, largely because of the long-standing United States commitment to the Library of Congress system and because of previous attempts by the Royal Society to formu- late a British standard." Neiswender gives as organizations interested in transliteration (in addition to LC), the British Standards Institution, the Board of Geographic Names (U.S.A.), the New York Public Library, the Slavic Review, and Mathematical Reviews. She also distinguishes transcription "as the process by which some idea is conveyed of pronounciation as well as of spelling from transliteration as the process by which characters in one alphabet are arbitrarily and unambiguously represented by characters in those of another, disregarding phonetic accuracy." 406 Finally we note recent ISO activities paralleling those of ASA for character recognition and codes for information interchange: "The ISO Technical Committee 97 . . . just concluded a series of meetings in New York City . . . World wide standards on optical and mag- netic ink character recognition and business ma- chines were drafted by the group . . . "The ISO draft proposal for magnetic ink char- acter recognition. Two different type fonts — El3BandCMC-7 — are being proposed. . . . "The new code (ISO) for information inter- change includes recommendations for both six-bit and seven-bits sets . . . Some of the code positions have been left unassigned to meet the needs of in- dividual countries . . . the code is generally com- patible with the American Standard Code for In- formation Interchange (ASCII) approved by the ASA last year." 407 3.9.4. Programs of Other Organizations An organization complementary in its special- ized field to the ISO is the International Electro- technical Commission (IEC), whose Technical Committee 53 is concerned with computers and information processing. The IEC recommends electrical standards only, whereas the ISO recom- mends international standards in all fields of en- deavor except electrical. The chairman of IEC/ TC 53 is A. B. Credle of the United States, who provided this information to the June 1963 Com- munications of the ACM : TC/53 held its first meeting in London in 1961. It has four subcom- mittees: SC 53A on digital input/output equip- ments, SC 53B on digital data transmission, SC 53C on analog equipments for information proc- essing systems, and SC 53D on input/output media. A joint steering committee has been set up, composed of the chairmen, secretaries, and technical advisors of IEC/TC 53, ISO/TC 97, and ISO/TC 95 (office machines), to avoid con- flicts of interest and duplication of effort. We note further that : "To facilitate the trans- lation of papers from one language into others, . . . the most important aspect is a precise, unified terminology. In particular, the use of interna- tional dictionaries is of the greatest importance. An example is the International Electrical Engi- neering Dictionary, which is now in its second edition and is presently being published in a third edition by the International Electrical Engineer- ing Commission." 40S By 1958 the International Federation for Docu- mentation had published a bibliography of stand- ards on documentation, with M. Schuchmann re- sponsible for the publication (1961 [505]). In 1961, with the financial assistance of UNESCO, FID published a supplement to the bibliography, bringing it up to date. Part I of the bibliography has a listing of the standards classified by coun- tries and ISO recommendations, supplemented by a list of all editions of the UDC recognized by FID; within the national groups, the standards are listed by serial number. In part II the stand- ards are listed in a systematic arrangement by UDC. Each standard is listed only once, under its main aspect. ECMA (the European Computer Manufactur- ers Association) came into being officially in May 1961 and has as its main purpose the development of methods and procedures for facilitating and standardizing the use of data processing systems. Information on the organization of ECMA was transmitted to the June 1963 Communications of the ACM by C. G. Holland-Martin, of Leo Com- puters, Ltd., of the Coordinating Committee, and D. Hekimi, the Secretary General. The technical tm Np;=-n-pn Schuller. 1960 [506], p. 379. 103 In the first TDCK test 100 queries were put to the two systems. Although the Uniterm system gave slightly hetter results, the UDC "procured many valuable documents not found by the other system, which fact will support TDCK in its con- viction that the two systems should be used together." In the second test 200 queries were put to the systems for retrieval of specific reports, known by title, from the center's collection of ASTIA-processed documents. The recommenda- tion resulting from these two tests is that a Uni- term system should be used for a collection of technical reports with known titles, for the follow- ing reasons: (1) there was only about a 1 percent chance that the report would not be discovered against 18y 2 percent by UDC; (2) the average search lasted only V/2 minutes against 10 minutes required by UDC; and (3) the librarian did "not need the help of a technical information officer to recover a specific report." 60 The third experiment was one of a series of comparative tests that have been conducted by Cleverdon and his associates under a grant to ASLIB by the National Science Foundation be- ginning in 1957. The tests were undertaken at the College of Aeronautics in Cranfield, England, and consisted of the indexing of 18,000 documents by the UDC, an alphabetical subject index, a special facet classification, and the Uniterm system of coordinate indexing. The first stage of the in- vestigation was concerned mainly with indexing the documents and preparing the four indexes; this stage was reported upon by Cleverdon in Sepember 1960 [122]. In the test program ques- tions were at first directed to a collection of 6.000 documents. These "Cranfield I" tests established the experimental recall ratio "by conducting searches for questions which had been based on documents known to be in the collection, and the result of this showed that, on an average, about 80 (percent) of the source documents were being retrieved." 61 Further tests against the full cor- pus of 18.000 items again showed a recall ratio ranging between 75 percent and 85 percent for all four indexing systems. 62 The investigators themselves have characterized the Cranfield I experiments as follows : "Its main endeavor was that it tried to measure the operating efficiency of complete indexing languages in a simulated real life situation, and also that it used, as its major measure of performance, a relatively crude gauge of relevance based on the retrieval or nonretrieval of a single source document." M The 1962 Cranfield _ Report (Cleverdon [124]) also emphasizes various reasons or failures to retrieve source documents. The findings have been sub- jected to some criticisms on these and other grounds, including apparent indecisiveness with respect to bases for choice among different systems. 60 Ibid, p. 3S6. 81 Cleverdon and Mills, 1063 [127], p. 6. G2 Cleverdon et al. 1964 [125], p. 87. 63 Cleverdon and Mills, 1963 [127], p. 12. The answer given by Cleverdon as to which of the systems should be recommended is that : "This is impossible to answer without qualification, for no system which has been investigated has shown itself to be so markedly superior as to justify its use in all conditions. The size of the collection, the number of users, obviously the subject matter; these are the type of considerations which would influence a decision." 64 Richmond, however, in a review of the Cranfield work, remarks that the answer as given by Clever- don "hardly satisfies the purpose of the experi- ment or does justice to its results." 65 Moreover, in the present state of the art of evaluation of infor- mation selection, storage, and retrieval systems, it may fairly be observed that: "The outstanding large-scale and realistic experimental work is that of Cleverdon," 66 and "many observors believe that the Cranfield study constitutes the most important work done in the field of cataloging in recent times." 67 Cleverdon and Mills have proposed a second series of tests called Cranfield II, which "will try to measure the impact on recall and relevance of particular indexing devices — the elements which go to make up a complete index language; and secondly, it will try to use decidedly more precise measure of relevance." In the Cranfield II series, 1,500 documents in the subject area of high-speed aerodynamics: "are being indexed with great thoroughness but without reference to any partic- ular scheme. For each report, all significant terms appearing in title, summary and text are given. . . ." e * Cleverdon and Mills are using four levels of description : an abstract, themes, concepts and terms for this study. 4.1.4. Problems of User Acceptance and Human Factors It should be recognized that the problems of user requirements and user recognition and accep- tance are, at least today, strongly interdependent. Several far-sighted and progressive systems have failed or have been discontinued because of lack of demonstrable, continuing usage value or because of user indifference. "No index . . . [and no cur- rent-awareness announcement service or search retrieval system] ... is useful unless it is used." 69 Further, "there is no virtue in any information service unless people use it actively." 70 The exemplary case is that the Chemical- Biological Coordination Center. While the pri- mary reason for failure was undoubtedly financial, it is likely that financing might have been readily attained had the CBCC's services been more widely known and used. It is to be noted that al- though at least 60 scientists had served on subcom- mittees of the CBCC, only about 30 letters « Cleverdon, 1962 [124], p. 92. 65 Richmond, 1963 [482], pp. 310-311. ^Koehen, 1963 [321], p. 12. 87 Randall, 1962 [467], p. 381. 88 Cleverdon and Mills, 1963 [127], pp. 12-13. 69 U.S. Senate, 1961 [541], p. 53. 70 Knox, 1962 [320], p. 27. 104 expressing regret were received at the time of the announcement of termination. 71 At the 1958 Western Reserve Conference on the possibilities of creating a "National Center for the Coordination of Scientific and Technical In- formation," Bernier remarked : "There have been at least three mechanized services available to scientists in general. All these three services had a deficiency of questions coming in to them. The Chemical Biological Coordination Center in Washington failed. The CBCC, the CNRS in Paris, and the Gmelin Institute in Frankfort re- port very few questions coming in to them. The reasons behind the failure of the CBCC and the shortage of questions coming to these three centers should certainly be investigated carefully, I feel, and explained before another general searching service is attempted." 72 An earlier and revealing example is that re- ported by Warheit as follows: "In 1951 Irma Wachtel adopted the Batten system to index the nuclear properties of the isotopes ... At the Ar- gonne National Laboratory sets of these cards were placed in the chemistry and physics libraries. Nothing happened. We then started pushing them and displaying them as if they were new books and succeeding in getting one person to use them. After two years of effort that is the meas- ure of our success in promoting this really useful indexing file. However, when the Nuclear Data Cards were issued by the National Research Council and these were placed in the libraries, we were immediately asked ... if they could have a set of the cards . . . The end product was an old- fashioned fixed file of over 2000 cards." 73 Studying the problems of centralization for documentation operations, the A. D. Little study- team reports : "Existing systems have not been used extensively, and our study shows that these systems have had difficulty in justifying their ex- istence in terms of their utilization." 74 More recently, Adkinson has reported as fol- lows : "For the systems listed in N onconventional Technical Information Systems in Current Use, 50 percent of those using punched card equipment answer fewer than one query per day; of those using computers, 50 percent handle fewer than three queries per day." 75 In the specific case of DuPont's Engineering Center, a 1962 statement reports: "Now in its fourth year of operation, EIC has an inventory of 10,000 accessions ... It is handling 1,200 inquiries a year, which is less than the rate should be." 76 There are many plausible reasons for neglect by users of new and improved services. One is a matter of inertia: "The most important step in the dissemination of scientific information is that of assuring easy access of this information to American scientists. Perhaps this should be the simplest task of all ; in reality it becomes the most difficult, not only because of logistics but because it involves overcoming the inertia of the American scientist and entering into competition for a por- tion of his time." 77 Further, "an outgoing, ag- gressive, use-oriented, service, in effect, asks the individual to change, sometimes in a radical man- ner, the information acquisition and handling pat- tern which he has developed for himself over his lifetime." 78 This problem is further aggravated by deficien- cies of training of scientists and engineers to make use of the traditional, much less the newer and non-conventional, types of services. "The train- ing of scientists to take advantage of the scientific literature of their own disciplines has been grossly inadequate, leading to ineffectiveness and in- efficiences that the nation can ill afford." 79 Thus, "a continuing educational program is required to convince all of today's engineers that modern documentation services can provide useful infor- mation that they are not likely to find themselves or even attempt to look for." so Again, "there is a tremendous education problem still ahead of us in terms of getting an engineer or scientist to recog- nize the existence of these resources and find ways to fit them into his daily work habits." 81 It should also be recognized that the problems of cooperation, compatibility, and convertibility must extend to the whole complex of documenta- tion operations in the handling and use of scientific and technical information. A five-minute, or a five-second, or even a five-millisecond response time to a scientist's search query will not avail much if, given the address-locators of possibly pertinent documents he must wait a frustrating 30 minutes or more for others to move about in the stacks in search and physical selection of the de- sired hard copies (many of which are likely to be "not-on-the-shelves"), or even if he must trans- port his microcopy "direct" retrievals to one or a few readers located elsewhere in his building, com- plex of buildings, or geographical area, and then there wait his turn in the queue of other clients. Sincere but often impractical gestures of collab- oration and cooperation have been typically bogged down by real differences of special-purpose interest, inertia, undue reliance on presumed customer acceptance of present products, and the all too prevalant human reluctance to change. ". . . In many instances there is opposition to cooperation, e.g., opposition based on fear that privately worked out documentation practices and forms will have to be abandoned in favor of a new, coordinated documentation. This is a psycho- logical problem." 82 n Dougherty. 1963 [1771. p. 14. " Shera et al.. 1958 T520]. pp. 204-205. "Wnrheit. 1956 [628], p. 270. " Arthur D. Little. Inc.. 1963 [3441, p. 45. 75 Adkinson, 1964 [6k p. 2. " Genereaux. 1962 [2141, P- 43. " U.S. House. 1959 [5871, p. 8. • 9 Knox. 1962 [3201. p. 27. '» Adkinson. 1962 [51. p. 50. 80 Genereaux. 1962 [214], p. 43. "U.S. House. 1963 [5821, p. 144, 83 Frank. 1959 [200], p. 502. statement of W. H. Carlson. 105 Probably far more serious from the standpoint of ultimately improving the utilization of recorded scientific and technical information is the fact that many present and potential users lack awareness of the availability of materials, services, informa- tion service products and guides to such products, services, or materials. This lack of awareness was strongly emphasized in the analysis of question- naires addressed to research administrators in the electronics industry with respect to their needs for current, on-going project information. In the first place, only a minority of the re- spondents emphasized the need for information prior to publication in their specific replies to the items of the questionnaire." Secondly, "a number of responses expressed quite general satisfaction with existing facilities and services in marked contrast to responses which were critical or which recommended radical new solutions." Thirdly, "a wide difference of opinion was noted with respect to the estimated numbers of people in the respond- ing organizations who need to be kept informed on the status of current research . . . ranging from one or a few to several thousand." Finally, and most importantly, "similar disparities in recogni- tion of the nature and urgency of the problem are perhaps implicit in a lack of awareness on the part of at least some respondents of what services are in fact now available. For example, one respond- ent complained that OTS facilities are not avail- able to him since he is not now a Government contractor whereas, of course, OTS services are available to the general public." a Corroborative evidence from the literature on the point of lack of awareness or lack of mecha- nisms for utilizing what is available may be further exemplified by the following: "There are defects in the existing information system, but there do not appear to be any major faults in the system as significant as the fact that so many who might use the system more, hardly use it at all." 84 Similarly, in a case of international cooperative enterprise, "O.E.E.C. entered this field a few years ago by launching a document exchange scheme . . . Reports of work issued by Government agencies should be exchanged between various national documentation centres . . . and through them made available to firms. This scheme has been only partially successful. Very little was ex- changed which was not received by the countries in any case, but when the situation was looked into a little closer, it often appeared that the material had not previously been easily available to industry because internal organization had not developed to meet the practical needs of the industrial user." 85 This problem is obviously related to that of realistic determination of user needs. "The es- sential difficulty is that, though the user may well know what he wants from an information service, he is in no position to know what he needs from it, namely what variations in the system would help most to further his work. Con- sequently, any action based on analysis of present user habits is unlikely to produce impressive results." 8e One of the further difficulties is that, "unfortunately, the problem is too often vaguely described as 'satisfying the user's need for infor- mation' without defining the need or appearing to recognize the extreme ranges of variation that can exist both in the nature of the user and the nature of the information." 87 The obvious further com- plication is that "the real needs of scientists and their conscious or expressed needs may be quite different. 88 Human factors, especially those involving local or special-purpose needs, also affect the content, order and format of bibliographic information that might be exchanged between cooperating orga- nizations. An example is given in a report of the American Petroleum Insitute's Central Abstract- ing Service. "Studies in one company of an abstract-writing style that drops bibliographical data to the end of the abstracts, to speed informa- tion transfer, had met with highly favorable reader reaction. This led to adoption of an in- termediate abstract- writing style that begins with document titles but places the rest of the biblio- graphical data after the text." 89 Such practices would obviously have a direct effect upon machine- usability of abstracts prepared in this form if they are to be interpolated into files containing abstract information in more conventional formats. Similarly, while the informality of corporate source entries at a Navy laboratory library — "We use nicknames (NOTS), abbreviations (lab., co., and est.), and colloquialisms (BuShips)," 90 — may be 'good public relations' for the local community, they would create obvious difficulties for interfiling or duplicate checking anywhere else. f In those areas of human factors involving ques- tions of user acceptance of machine products, much basic fact-finding remains to be done. For example, KWIC indexes and other indexes com- piled or generated by computer are undoubtedly serving genuine current-awareness needs and they can be prepared promptly and cheaply. However, they are marred by disadvantages of line length, arbitrary truncations of words and excessively long blocks of entries centered on some indexing words. The limited-to-upper-case types of computer printout typically offered afford disadvantages of acceptability and probably of physical legibility in terms of the quick scanning that is desirable. Within the general problem areas, then, it may be concluded that "Human factors and machine factors constitute the two major operational con- s' U.S. Senate, 1961 [590],, pp. 108-109. 84 Urquhart, 1952 [604], p. 234. 85 King, 1955 [316], p. 9. 88 Bernal, 1959 [54], p. 79. 87 Hillier, 1962 [261], pp. 31-32. 88 Brownson, 1962 [84], p. 77. 89 Weil et al., 1961 [635], p. 58. oo Carlson, 1952 [97], pp. 121-122. 106 siderations that must be systematically examined, and which will lead to a host of research prob- lems." 91 Before considering some of the specific problems introduced by prospects of mechaniza- tion, however, we shall first examine some of the special problems which have arisen in connection with changes in the nature of materials to be handled. 4.2. Special Problems Raised by Changes in the Nature of the Materials to be Handled Major changes in the nature of the materials to be handled in scientific and technical information systems raise special problems. These major changes include the phenomenal growth in the amount of the literature to be covered, the in- creased quantity and importance of foreign lan- guage materials, the emergence of less conventional forms of publication and communication as bear- ers of forefront information, and the use of non- documentary forms, especially in the area of photographic and sound recordings. "To the documentalist, information resources consist not only of the formally organized contents of libraries and archives — their books, journals, serials of all descriptions, reports, all the tradi- tional library materials, whether in original or reproduced form, but also of stored data in hundreds of forms: reprint and abstract card- files; graphic charts, such as infrared spectra; tabular records of experimental data; data stored on IBM and edge-notched cards, on microfilm, on recording tape, on electronic memory drums, et cetera." 92 As an example of new and specialized types of material for which bibliographic controls, an- nouncement mechanisms and means for exchange or loan are needed, the case of the technical film may be considered. "The technical film is becom- ing of increasing importance in technical informa- tion dissemination." 93 These films may, for ex- ample, provide data on phenomena not detectable by the human eye and involving slow motion, time lapse, or photography in invisible parts of the spectrum. "The Encyclopedia Ginematographica ... is a growing collection of research films initiated . . . by Dr. Gotthard Wolf, of Gottingen, West Ger- many. Operated on a supranational basis, the program involves the collection, classification, and publication of carefully selected research films that present motion phenomena of scientific values .... The encyclopedia now consists of 340 separate film units . . . " 94 Another example is "most interesting among new Library acquisitions will be the Psychiatric Recording Library . . . recordings ... of spon- taneous group discussions among psychiatric patients . . " 95 Further, "for biomedical information recorded in audiovisual form, the National Audiovisual Facility of the Communicable Disease Center of the PHS should be developed to the point where it is analogous to the National Library of Medicine as a central resource for such records and a com- piler of 'tools' for their retrieval." 96 Unusual materials in the sense of traditional documentation may extend also, as in earlier days at the U.S. Patent Office and the Army Ordnance Corps Museum, to models, devices, and physical samples. Thus, as a result of the post- World War II foreign acquisitions program, "through OTS and Army, American industry has been able to borrow for testing and examination, many fasci- nating and novel devices: a horizontal mercury type chlorine cell, a miniature Diesel locomotive, a portable mortising tool . . . ." 97 At an ASLIB Conference held in London on the prospects for cooperative action in the field of technical translation, the Chairman opened the sessions with comments on the major problems of the information specialist, first, that he must scan an enormous quantity and variety of materials, and second, that much of what he must scan is written in various languages. ". . . Much of the output of information is to be found in technical periodicals and in the flood of ad hoc reports which tumble forth in increasing volume from Government departments and other official agencies, here and abroad. The organiza- tion of the publication and issuance of all this material is as nearly chaotic as any human activity could be: we all know that for every significant article in any field, published in the dozen or so periodicals in which it could be expected, there are likely to be at least one or two equally sig- nificant articles in the remaining tens of thousands of regular periodicals, to say nothing of the occa- sional and non-serial publications. This points to the information worker's first problem — the scan- ning of a vast and growing body of literature for material relevant to the needs of those he serves ... In doing so, he meets his second main problem : only 40 percent are in English. The re- maining 60 percent are in a foreign language — or, rather, in many foreign languages." 98 Problems of coverage and scope in terms of cooperative attempts to attack these problems in the light of the phenomenal growth in the sheer quantity of items to be considered are stressed throughout the pertinent literature. Two further observations may highlight these difficulties here. Thus, "the UNESCO Scientific Abstracting Con- ference Report in 1949 accepted one-third as the n Stanford Research Institute, 1963 [542], p. 9. 02 Adams, 1956 [3], p. 73. 83 U.S. Senate. 1960 [592], p. 48. "Scientific Information Notes 2, No. 5, p. 8 (1960). M Army Medical Library News, 1, 3 (Nov. 1945). M "Communication Problems in Biomedical Research," 1963 [403], p. 16. "'"Distribution," 1947 [175], p. 10. " L. V. Chilton, statement in "Technical Translation : Co- operative Action," 1960 [566], p. 129. 107 fraction of scientific papers published which are abstracted in some way," " and the situation has improved only a little: "Some indication of how far we are from the indexing coverage of the world's literature may be obtained by noting that in 1961, all of the major U.S. scientific abstracting and indexing services combined . . . will cover only about 16,000 of the world's estimated 30,000 scientific and technical journals . . ." 10 ° One of the most dramatic changes in the nature of the materials to be handled in systems for proc- essing scientific and technical information has been the emergence of the unpublished technical report as the major vehicle for communicating current research and development progress. The flood of such reports has resulted in a number of special problems and difficulties, ranging from the inadequacies of conventional coverage to specific questions of specific identification. The nature of the first major problem, that of inadequate cover- age by libraries and the traditional abstracting and indexing services, has been stressed in a study by Herner and Herner. For example, "with the exception of Nuclear Science Abstracts, the major abstracting and indexing services in the United States have tended to ignore the reports that lie outside well defined and easily identifiable series . . ." 101 The reasons include small volume publication, the relatively ephemeral nature of the subject matter covered, and quality considerations. A second major problem in the handling of the report literature is the question of what in tradi- tional systems is considered the "main entry" for cataloging purposes. The decision that in many cases the corporate source should logically consti- tute the main entry has complicated the already controversial area of rules for descriptive catalog- ing. As we have seen, the earliest attempts to achieve interagency compatibility in the U.S. Government had as a principal concern the ques- tion of proper recording of corporate authorship. Writing in 1952 on the problems of identification and control of the report literature, Warheit in- dicates that coverage desired at AEC ranged from omission of all authors with recognition only of the divisional director in a laboratory to listing all authors, all who worked on the research pro- gram, "plus the total hierarchy who approved the report and were generally responsible for the work done." 102 The argument advanced by Taube (1950 [551]) with respect to corporate author entries, that the cataloger should not be required to go beyond the information available to him from the item itself, becomes all the more cogent when the human cataloger is replaced by machine. The argument will also extend to other entries — the dates of births and deaths of authors, for example. True, m Visscher, 1954 [617], p. 85. 100 Bourne, 1962 [72], p. 162. '»i Herner and Herner, 1959 [254], p. 192. Sae also p. 29ff. of this report. im Warheit, 1952 [627], p. 105. the machine can apply rules consistently, but the costs of compilingj maintaining, and updating the necessary lookup lists, directories, biographic data files and the like would add prohibitive extra expense. Special problems are also raised with respect to the control of the report literature, especially in terms of the prospects for mechanizations. Exist- ing announcement bulletins such as TAB or USGRR often carry items which have no title, but where the input to the bulletin-preparation is in form suitable for machine-compilation of title in- dexes, suitable provision must be made for the sup- ply of a pseudotitle which the machine program can use. Another specific problem is that of identification of reports by one or more numbering systems. Warheit points out that "although potentially it is one of the best and most convenient devices for identifying reports, the report number in actuality does have many drawbacks. To begin with, many originators do not assign numbers to their docu- ments and, with each recipient providing his own distinctive number to these 'orphans,' there has developed a veritable babel of identifying symbols and notations." 103 Connor suggests that along with other specifica- tions in government contracts there be "a brief statement setting forth physical and bibliographic requisites in the preparation and publication of reports," including a reasonable scheme for num- bering. 104 Rubinoff of the Technical Operations Committee, Information Retrieval Subcommittee, of the Institute of Electrical and Electronics Engi- neers (IEEE) declares that the IEEE, among other tasks, needs to agree on a "document numbering system for all documents including Standards." 105 Mahany, Creager, and Herner have reported to the National Science Foundation the results of their study on functional symbols currently in use in government reports (1962 [367]). The study suggests "how widely agencies differ in their re- porting policies generally, as well as in the names and symbols they apply to constituent parts of their report systems." The semantic problem — that of names and symbols — is considered to be "but one of a series of problems, some wider and more fundamental, in report production and con- trol." Until resolved, the differences in reportage among the agencies "must continue to bar the standardization of report designators, unless sim- plified designators are designed with greatly cir- cumscribed functions." 106 The design of a stand- ard Government or Department of Defense-wide system of report designation would presuppose, and its implementation require, extensive revision of most agencies' reporting practices both in scope and structure. Any system developed would de- nk Warheit, 19552 [627], p. 106. 104 Connor. 1956 [134], p. 155. 105 Rubinoff, 1963 [500], p. 3. 106 Mahany et al„ 1962 [367], pp. 29-30. 108 fine a standard code of symbols serving to give each report title a unique identity and yet relating it to a report series and a report system and pos- sibly providing other information. It would probably require essentially complete change by all agencies using such report designators. Unpublished conference or symposium papers, as well as published proceedings of such meetings, are among the newer, important forms of informa- tion sources. "Conference papers, conference proceedings, and information about the confer- ences themselves represent a serious problem in the technical information field. Papers presented at a conference may contain the most up-to-date information available on a subject, and as a conse- quence considerable effort is devoted to getting copies of them. At the same time such papers are frequently difficult to locate." 107 Experience in support of this opinion has been reported : "Over two-thirds of the papers published in the reports of ten conferences were not individually abstract- ed in the journals searched, and nearly half were neither abstracted nor listed and indexed." 108 Suggestions have been made for lessening such problems : "If all conferences adopted a brief di- rect title incorporating the subject of the conference and left all the attributions of orga- nizations, sponsorship, etc., to a subordinate por- tion of the title, the cataloguer's and the reader's difficulties would be greatly reduced." 109 At the 1958 International Conference on Scien- tific Information Liebesny reported [342] on "Lost Information; Unpublished Conference Papers." Several years later Ruth Hooker (1962 [271]) re- ported that proceedings of conferences and sym- posia were becoming an increasingly important source of information at the Library of the Naval Research Laboratory, and she raised the question as to whether material so published is adequately covered in abstracting and indexing services. Some of the government report announcement bulletins attempt to cover this material. Thus, in each issue Nuclear Science Abstracts announces its methods for (1) arrangement of abstracts and (2) supplying indexes to the publication. Part of the announcement regarding abstract arrangement states, "In the case of comprehensive progress re- ports, or conference proceedings and symposia, an abstract is provided both for the whole publication and for each individual topic, sentence, or paper." An example is the announcement of three volumes of the Proceedings of the Seminar on the Physics of Past and Intermediate Reactors, held in Vienna Austria, August 3-11, 1961. Volume I is an- nounced as item 26503, followed by separate ab- stracts for items 26504 through 26534 on neutron physics and integral experiments in that volume. In 1963, NSA began a concerted effort to account for individual papers associated with a particular conference not yet published in formal proceedings. The conference is assigned a CONF — designation, and thereafter, all papers associ- ated with this conference are so identified. The identifying numbers appear in the abstract, and also in the report number indexes, grouped to- gether. Availability of the individual papers is shown at this point. If proceedings are published later, the availability of the proceedings as a whole is shown and the individual papers are dropped. STAR has a method generally comparable to that of NSA for announcing the papers given at a meeting. "Availability of Nuclear Science Conference Literature (TID-19000) is a new service initiated by DTIE [The Division of Technical Information Extension of the Atomic Energy Commission] in July of this year (1963) which should be a partial solution to the problem. It will give information about the availability of proceedings or individual papers of conferences in which DTIE is interested and in which other AEC installations have by their queries indicated an interest . . . The avail- ability information will include such items as fu- ture publication plans, negative publication plans (much time can be wasted looking for proceedings that will never be published), journal citations, (proceedings are particularly hard to locate if published as part of a journal), and known inten- tions of anyone to translate." 110 On an international scale also, the question of conference papers and proceedings is being studied : "The International Federation for Docu- mentation (FID) on 14 September 1959 signed a contract with UNESCO ... for the preparation of a 'study on the content, influence, availability and value of scientific conference papers and proceedings'." iai The Union of International As- socations, Brussels, Belgium, has received a Na- tional Science Foundation grant "to compile and publish a monthly bibliography of all proceedings and reports resulting from international meetings, congresses, conferences, and symposia." 112 From another point of view, "ISO TC/46 is now study- ing a preliminary draft international recommenda- tion on title pages, which includes a special paragraph on symposia." 113 Finally we should note, in terms of the problems implied by potential exchange, interfiling and compatibility among information systems, the question of the handling of "separates," both pre- prints and reprints. Brownson reports a recom- mendation of the Royal Society Conference (1948[498]) with respect to desired common practice: "It is recommended that all separates (reprints) be headed with a full reference (name, number, volume, and date of the original journal), and keep the original pagination . » 114 1OT Davis. 1963 [158], p. 237. 108 Hanson and .Tones, 15)61 [236], p. 145. lw Jolley. 1963 [303], p. 60. 110 Davis, 1963 [153], p. 237. 111 Report by Polndron, published in two parts UNESCO Bull Libs. XVI. 113-126 and 165-176 (1962) [459], [460]). 112 Scientific Information Notes 3, 6. 11 (1961-2) " 3 UNESCO Bull. Libs. XVI. 172 (1962). 1,1 Brownson, 1952 [85], p. 38. 109 Again, with respect to questions of descriptive cataloging alone, for future compatibility in the case of preprints, what date shall be used, that of the preprint itself, if ascertainable, or that of sub- sequent publication ? 4.3. Special Problems Raised by the Prospects for Mechanization One of the principal reasons for increasing con- cern about cooperation, convertibility , and com- patibility is precisely the potential impact of machine processing capabilities. Mechanized ca- pabilities for (a) processing bibliographic data for multipurpose use, (b) preparing tools for current- awareness searching and dissemination, and (c) providing techniques for storage, retrospective search, and retrieval create new needs and provide new opportunities. First, machine use — and in particular, prepara- tion of material for such machine use — are expen- sive operations. The advantages that machine techniques offer (of speed, consistency, multiple copy possibilities, storage economies, direct tie-in to various communication links, and rapid updat- ing of files, catalogs, and indexes, and the like) can be best achieved, despite the costs, if there are many users. Similarly, many contributors may be required. Cooperaton among users, among con- tributors, and between them and the processing centers involves mutual agreement as to coverage, format, division of effort, priorities, and proper balancing of service in terms of the various customer requirements. Second, machine use requires material in ma- chine-usable form. Once material (whether it is a catalog card or other form of descriptive cat- aloging information, a record of index entries, an abstract, or even full text) is in machine-usable form, it can be reused for many different purposes and it can be replicated for use in other locations and by other organizations. Thus exciting new possibilities are opened up not only for the ex- change of records and other materials but for sig- nificant reduction in duplications of effort such as are involved in the cataloging and indexing in one agency of an item already cataloged and indexed by others. However, realization of these poten- tialities depends upon compatibility, or at least convertibility, between systems with regard to such varied factors as the physical characteristics of recording media, the format of recordings, ma- chine language and character sets used, the identi- fication of contents of recorded information, the special symbols or codes used, the citation or cata- loging practice, and many others. Third, machine use and communication system links imply new organizational and service possi- bilities. For example, many different users may tap in to a remotely located source of information about literature, information, and data availabil- 110 ity. Cooperating centers can maintain their indi- vidual responsibilities at their separate locations with effective coordination at the national level. In practice, however, mechanization of systems of documentation and information handling opera- tions has to date been limited to relatively small, specialized collections, such as systems for han- dling less than 25,000 internally generated tech- nical reports for which no conventional means for control had previously been attempted. There are a few exceptions, such as the very large Mini- card and IntelloFax installations in military and intelligence agencies. Studies made in recent years by Stanford Re- search Institute, Arthur D. Little, Inc., the Library of Congress automation team, and a group studying the possibilities of a cooperative library system to be supported by local industries and Harvey Mudd College provide a consensus as to the limited applications of machines to date and as to the lack of compatibility both functionally and with respect to equipment used, for example : (1) "No present machine is equal to the task of storing the immense quantities of information available. Even further from solution is the prob- blem of achieving rapid access to a specific small fraction of a very large body of stored informa- tion." 115 (2) "Present automation techniques do not deal adequately with raw text. At the outset only catalogs, inventory files, and indexes should be considered for automation. . . . The economic and technical feasibility of automatically retrieving information directly from the text of documents has not yet been established and is an extremely complex subject. For this reason, there is likely to be greater emphasis for some years on the re- trieval of bibliographic information." 116 (3) "The investigation revealed that there is no standard method of handling information storage or retrieval. Each organization tries to meet its own needs by its own methods. Just as there is no standard procedure for information re- trieval, so there is no standard equipment. All systems studied share the problems created by the tremendous amount being published and the time required to prepare it for processing through auto- matic equipment. All lack adequate storage (un- less a large computer is used) and all lack a rapid printing process to handle the output . . . ." 117 For these reasons, there is as yet no discernible "state of the art" with respect to cooperation, com- patibility, and convertibility, even as between a few mechanized systems, much less between large and small, mechanized and nonmechanized, con- ventional and nonconventional. Instead, what can be discerned is a state of affairs, from which two contradictory conclusions emerge. One is that while maximum cooperation and convertibil- 115 Stanford Research Institute, 1963 [542], p. 10. "'King, 1963 [318], p. 11. 117 "A Joint College-Industry Library With Automata," 1964 [302], pp. 13-14. ity must be urgently encouraged, most efforts to achieve compatibility or standardization are pre- mature. The other is that, whether improved cooperation and convertibility are or are not achieved, greater compatibility is imperative now. Still other, but related, reasons for concern, ex- pressed in the literature to date, include the following : "The trend toward mechanization con- tinues throughout major Federal agencies. This trend makes possible, indeed it renders essential, efforts to assure reasonable compatibility between the machine systems." 118 "The danger in the current situation is that attention will be so strongly focused on automa- tion that the primary problem of achieving com- patibility in the scope and use of agency collections will be neglected." 119 Writing about the problems of planning for compatibility within a single large company, but with cogent comments much more generally appli- cable, Anderson (1962 [29]) provides a detailed and thought-provoking exhibit of discussion ques- tions which should be investigated in the develop- ment of compatible systems, ranging from what identification criteria are needed and whether dif- ferent ones required can be merged into a single system to the possible pitfalls of coding and fixed fields for print-outs, conversions, sorting, and the like. She stresses that : "It is important to remem- ber that attractive but irreversible decisions may beckon at every turn. . . . Such decisions should be made with cautious deliberation where costly operations, such as the analysis of documents by subject specialists, are involved." 12 ° She con- cludes that "compatibility with the future should not be sacrificed for compatibility with all parts of the company." 121 A major reason for current concern is that mech- anization of some sort is now in progress or is contemplated for the near future in many differ- ent organizations, with respect to quite different areas of machine application, and with respect to collections ranging from quite small to the very large. Decisions at a variety of levels in a variety of places, and intending to accommodate, locally, a wide diversity of interests, needs and facilities can not only contribute to but escalate present in- compatibilities to a level of continuing chaos. This is because of the sheer costs and time required to convert any existing records and items to machine- usable form. 4.3.1. Machine Language and Bibliographic Information Machine language compatibility tomorrow im- plies the requirement for considering such com- patibility today in present systems as they are cur- rently used. Achieving compatibility in the rou- tine steps to which mechanization has been prin- cipally applied to date should be easier from many " s Cahn, 1962 [94], p. 23. ""Crawford Report, 1962 [465], p. 35. ""Anderson, 1962 [29], p. 114. 121 Ibid. standpoints than that which would be required for those operations and processes so far still re- quiring human analysis and judgment. Why cannot standards for the content and format of catalog cards be applied? Why cannot standard conventions for descriptive cataloging be devel- oped and adopted? Why cannot standard meth- ods for recording, preparing and reproducing catalog cards be established as is being explored for the Library of Congress by Buckland and as- sociates of Inforonics, Inc. (1964 [144] ) ? Prior efforts to achieve agreement on rules and standards for descriptive cataloging are reviewed elsewhere in this report, but it is probably impor- tant to emphasize again that while the pessimism expressed by Wright in 1956 . . . "When the sub- ject for this paper was suggested . . . my first reac- tion was to protest that a report on progress in the revision of our catalog code would resemble only a dissertation on the religion of Frederick the Great or a monograph on snakes in Ireland — a collection of blank pages" 122 — is probably not as justified today, there is evidence that many very real difficulties remain. Permissive alternatives in cataloging practice, which appear to be a major result of cooperative attempts toward standardization agreements to date, merely complicate the mechanization prob- lem. For example, deliberately cooperative at- tempts to achieve greater compatibility often lead to greater difficulties of compatibility and convert- ibility on the practical side. Thus, while "the possible disadvantage of bilingualism (in a pro- posed single international Journal of Physics Ab- stracting) seemed a cheap and acceptable mutual concession," 123 it is also clear that the "constant recommendation of alternatives may be a rational procedure for securing agreement in an interna- tional meeting but in itself it weakens the preci- sion, the economy, and the efficiency of the catalogue." 124 From the standpoint of prospective mechaniza- tion, moreover, the problems are aggravated not only with regard to substantive questions but with respect to details of storage requirements, typog- raphy and character sets as well. (Parentheti- cally, the physical medium for storage of catalog entries was a factor in early Anglo-American co- operative efforts when "by adopting centimeter- sized cards, American librarians gave evidence of their hopes for international exchange of catalog cards.") 125 Decisions on elements to be included in records and files, the coding or notation systems adopted, and the number of types of characters to be rec- orded will have a direct and obvious effect on pres- ent and future storage requirements. The impli- cations of potential mechanization on the format and content elements of bibliographic references > 22 Wright, 1956 [655], p. 331. 123 Bontry, 1959 [74], p. 504. 1M .Tolley, 1963 [3031. p. 61. «s Ludington, 1954 [351], p. 194. Ill and citations are particularly evident in the areas of interfiling catalog card information for ma- chine search, of producing machine-compiled cita- tion indexes, and of machine processing of full text. Margaret Thompson (1963 [570]), describes a computer program that will take bibliographic re- ferences or citations, once they have been com- piled, and process them automatically, arranging the citations into a standard format so that the parts are unambiguously identified. Nine items, the basic elements of a descriptive catalog entry, are accommodated in her program: (1) author or authors, including the designations "Jr." or "III ;" (2) title of book or paper; (3) title of journal ; (4) volume, (5) issue, and (6) page numbers; (7) name of publisher, and (8) city; and (9) year of publication. The program also can sort and match the parts of the citations, thereby making pos- sible the identification of similar citations and the effecting of various rearrangements of items. We note, however, that the citations must first be compiled and suitable designators for the ele- ments given. The difficulties here are at least two- fold: first, that the standardization of rules has not yet been achieved, and second, that prac- tices are not consistent even for humans and are therefore not yet amenable to direct processing. Clapp remarks on the somewhat ironical situation that ". . . While in the traditional systems the identification of a document (known in library parlance as descriptive cataloging) is considered perhaps less difficult than content analysis (known as subject cataloging), with the computer the re- verse is true." 126 With respect to machine compilation of citation indexes, for example, Pauline Atherton points out that: "Unless the references are assembled in a standard form, the data collected in separate cita- tion indexing operations can not be matched and resorted by machine. A rigidly standard form for machine manipulation is necessary to insure this type of coordinated searching and sorting. To date almost everyone developing citation indexes has their own code and form." 127 What might be called "chaos in card catalogs" is of course as serious in many manual systems as for those involving potential mechanization. For example, in the 1944 survey of the Army Medical Library, it was found that : "Two different systems of filing have been used, one, a word-by- word sys- tem, and the other, a letter-by-letter system. As a consequence identical entries may be filed in two different places . . . "Even when an entry has been found in the card catalog . . . the next problem is to find a call number or class mark on the card . . . The most prominent notation on many cards is the accession number, but that is of no use for this purpose. Another number, the one in the upper left-hand corner, has to be disregarded also, since it is an indication of how many cards were typed for that particular item. Other figures in the same corner may, however, be location symbols. In blue or black pencil, or typed on the card, they may indi- cate the pamphlet box where the item is filed. Sometimes that number is prefixed by a symbol such as "P" or "Bx". . . the box number may have been printed at the end of the entry . . . Disserta- tions are indicated by a star, prefixed to the first word of the title . . ." 128 and so on. Hopefully, the example is of academic interest only by now, but it serves to point up the very real difficulties of transcribing catalog cards. On the other hand, "... computer people are learning somewhat to their amazement (and the librarians' amusement) that, as one 'machine' man put it recently, 'every mark, every space, every position, every word on a Library of Congress catalog card means something.' " 129 . Under a grant from the Council on Library Re- sources, Inforonics and the Library of Congress are studying possibilities for multi-purpose re- cording of bibliographic information in machine- usable form. A press release sets forth the study objectives, in part, as follows: "The firm . . . has undertaken to demonstrate that bibliographic in- formation, when suitably punched into a perfo- rated paper tape record by a tape-producing typewriter, can be reproduced in any form, whether complete or abridged, typewritten or printed in any one of a variety of type faces, as may be required for the various records used in library operations. The purpose of the demon- stration is to lay the basis for seeking general agreement among libraries and other users and processors of bibliographic information as to the standards of conversion of such information to machine-readable form in order to assure com- patibility and interchangeability of the product resulting from work performed at different locations." 13 ° Supposing that such compatibility can be achieved for the future, however, there remain the enormous burden and cost of integrating the new products with the old. Here a number of special problems and factors meet to emphasize the pres- ent seriousness of the impact of proposed mechan- ization. First, there is the prospect that automation of the more routine aspects of catalog- ing and bibliographic work will proceed well in advance of the design of systems capable of mechanized search and retrieval for the very large collections. Second, there is the tremendous problem of converting present catalog files and related bibliographic information into machine- usable form. Third, there is the question of character sets required both for input and output. 128 Clapp. 1963 [116], p. 7. 121 Atherton, 1962 [39], p. 4. 1= s MPtcalf et al., 1944 [380], p. 23-24. 129 Adkinson, 1964 [6], p. 5. 130 Council on Library Resources, 1964 [144], p. 1. 112 With respect to the first of these problems, the team studying possibilities of automation in the Library of Congress has concluded that "the re- trieval of the intellectual content of books by auto- matic methods is not now feasible for large collec- tions, but progress in that direction will be ad- vanced by effective automation of cataloging and indexing functions." 131 Their report stresses, however, that ''there must be compatibility be- tween the Library of Congress and other research libraries, particularly with respect to the kinds of equipment developed. There is some urgency with respect to this compatibility since some efforts at automating individual libraries are already under- way. The effort of establishing compatibility, or at least appropriate interfaces, is probably small compared to expected benefits." 132 The second major point is more general in that it applies to all information that is required to be in machine-readable form and this problem is therefore separately discussed. The question of the number of characters in catalog information, however, has been investigated in enough detail to provide data of both present and future concern. First, with respect to the number of characters per record, Sprenkle and Kilgour have made prelimi- nary investigations of the average number of units (characters, punctuation marks, and spaces) in main entries, titles, imprints, notes, collations, subject headings, and added entries of sample cata- log cards used in several biomedical libraries. They found, for example, that the number of such units used for personal author names varied from 10 to 37. In general, their results showed consider- able uniformity of the average number of charac- ters used. They conclude : "The internal con- sistency of the results of this study strongly justi- fies this type of analysis for producing findings useful in estimating machine capacities required for catalogue cards." 133 The conclusions of the Library of Congress auto- mation study team 134 are that : "The conversion of catalog cards is an enormous task; the National Union Catalog, for example, contains about 15 million cards with an average of approximately 50 words per card. Running text can be key- punched and verified at roughly one cent per word. At this rate the complete conversion of the Na- tional Union Catalog would cost about $7.5 mil- lion. However, the heavy mixture of numerals and the necessity for a certain amount of format- ting, error correction, and quality control suggest that the one-cent-per-word figure may be some- what low." 135 4.3.2. Storage Requirements Another special problem raised by present and future prospects for the mechanization of docu- ia Kins. 1963 [318], p. 2. "» Ibid., p. 18. w» Sprenkle and Kilgour, 19C.3 [541], p. 205. "* King, 1963 [318], p. 9. «* Ibid. mentation operations is that of storage capacity required. In September 1963, John Senders re- ported on three different calculations of informa- tion storage requirements for the contents of the libraries of the world. He summarizes : "The total range of the various estimates is only slightly more than one order of magnitude : 7.5 X 10 7 to 7.7 X 10 8 . At 10 5 words per book and five letters per word, the number of letters to be stored is between 3.8 X10 13 and 3.8 X10 14 . At 50 characters per alphabet and with 50 alphabets, there are 2,500 characters to be identified, or about 12 bits per character. The information storage measure of the "world's literature" then is between 4.6 X10 1 * and 4.6 X 15 bits; 2X10 15 is conveniently close to the mean . . . "The growth rate is estimated to be about 3.1 percent per year, doubling in 22 years. Thus, the current addition rate is about 6.2 X10 13 bits per year, or 2 X 10 6 bits per second." 136 Even if mechanized records of library and in- formation center holdings are to be limited for some time to come merely to the contents of sec- ondary records, such as present-day card catalog information, the storage and machine processing problems remain enormous. The King Report estimates, for example, that "the procurement of hardware, software, and necessary file conversion for automation of the central bibliographic opera- tion of the Library of Congress would amount to about $30 million.'" 13? The situation of the combined departmental li- braries of the Federal Libraries is not dissimilar. "As of June 1959, the holdings of the 212 refer- ence libraries reporting in the Brookings survey totaled almost 14 million books, serial volumes, and uncataloged pamphlets (to which should be added some 5.6 million technical reports held by 107 of the libraries) ." 138 Presumptively this total would include a large proportion of overlap in holdings, but how can the specific duplications within a group of diversified collections of this size possibly be determined without machine processing ? Related to the general problems of machine stor- age requirements are the more specific questions of fixed, specified, and variable record lengths al- lowed for records, fields, words, and other entries, and questions of possible truncations or artificial transformations used to reduce the redundancy of natural language words and texts. Patrick and Black give an example of special machine considerations involving the need for fur- ther agreements on descriptive cataloging conven- tions and practices. ". . . If we held 2 full deci- mal dates in 6-bit BCD (binary coded decimal), 48 bits would be required to store the information about the year of an author's birth and the } T ear of his death. By adopting a suitable convention (the base year 1000) and an appropriate pair of defini- "" Senders. 1963 [5151, P- 1068. >« King, 1963 [318], p. 2. »» Evans, 1963 [18S], p. 3. 113 tions — the first field contains an increment such that the year of birth is obtained by adding the first field to the base year, and the second field contains age at death — we can reduce the number of binary positions required from 48 to 17 . . . Whenever information is required on a printout ... it will be displayed as it now appears on a catalog card." 139 These authors suggest further that ". . . the li- brary community should be stimulated to debate, in publication and open forum, the requirements for the master catalog file. In particular, the com- munity should be encouraged to discuss, and even- tually to agree on, the following: 'Resolved, the present catalog card contains information deemed unnecessary in an automated system for reasons of economy. These items are . . .' Some of the topics to be discussed are the myriad type sizes, fonts, and faces that have been used on catalog cards." 14 ° Similarly, Wright suggested earlier: "We should start with the question, Why do we need the author's name ?' And when we have answered that we should consider whether the answer im- plies the need for the full name and dates of birth and death. If it does not we next should ask what other reasons there are for including them. Are these reasons sufficiently cogent to justify the cost of securing and recording the informa- tion?" 141 The problems and costs of securing and record- ing such information raise the further questions of the language and media of the documentary lan- guages themselves ; the language and media of en- coding, inscription and storage of document sur- rogates and condensed representations used for search, selection and retrieval, and the language and media of reproduction, display or transmis- sion. We shall next consider some illustrative problems of input, internal machine language, and character sets. 4.3.3. Problems of Input First is the question of obtaining machine- usable input to any mechanized system. Modern machines are capable of working with letters of the alphabet and numbers when they are expressed as electrical signals or other physical states rep- resenting those alphabetical letters or numbers. This physically encoded expression of the alpha- numeric characters is called "machine language." The machine languages require the automatic or manual transcription of the conventional alpha- numeric symbols into a coded form which may or may not be legible to the human eye. The coded form may be a specific pattern of holes in one column of a punched card; other forms of coding may be utilized in punched tape, edge- notched cards, optical spots, or magnetic cards, tape, or drums. 139 Patrick and Black, 1964 [450]-, p. 46. 140 Patrick and Black, 1964 [450], p. 48. '"Wright, 1956 [655], p. 33. To elate, there are three principal ways to ob- tain or record information in machine language form: (1) actual language or by-product record- ing at point of origin as in tape typewriters or typesetting paper tape; (2) manual keyboard operations; or (3) automatic character recogni- tion, not yet developed to the point of feasibilty for general purpose text reading. This machine- usable text may be regarded as the representation of the alphanumeric symbols constituting a mes- sage or messages in the machine languages directly usable both in specific input-output devices and media and in the internal memory and processing units of computers and similar data processing equipment. Material can be considered machine usable whenever it may be processed directly by machine without manual intervention, even though it may be automatically converted from one ma- chine language to another during processing. The first method is exemplified by the Inf oronics studies previously mentioned, by new develop- ments in computer-related automatic type-compos- ing equipment such as GRACE in the MEDLARS System, and by the availability of paper tapes used to drive more conventional typesetting equip- ment, such as Monotype. Up to the present, how- ever, the potential attraction of the availability of machine-usable text in the form of Monotype tape has been almost universally disappointed by the lack of suitable conversion equipment to rerecord the information given in a 30-column layout to 6-, 7-, or 8-channel tape usable by computer, or to punched cards or to magnetic tape. Another dis- tinct problem is that of detecting corrections that have been made and handwritten indicia for changes of type style, font, and the like. Some text is already available, whether obtained by this method or by manual keystroking opera- tions with a tape typewriter or a keypunch for punched cards. The National Science Foundation announced in 1963 the awarding of a contract to Thompson Ramo Wooldridge, for a feasibility study of a "center for text in machine-usable form," the center to collect and make available machine-usable texts for documentation re- search. 142 A report as of February 1964 (Mersel and Smith [379] ) gives a total figure of approxi- mately 50,000,000 words of text so available, but this includes a large proportion of nonscientific text, such as newspaper and popular magazine materials. In terms, then, of material now available and the fact that the new type-composing equipment de- velopments can operate only on future materials, not the present store, the principal method for preparing machine-readable material remains that of manual keyboard operation. In this second method of preparing machine- usable input, the typist or keypunch operator transcribes indicated copy into a desired machine 142 See Scientific Information Notes for October-November 1963, p. 1. 114 language form on punched cards, on punched pa- per tape, or directly to magnetic tape. The under- lying difficulties here are time, cost, and availa- bility of skilled operators. "Punched cards must be produced manually from data in the source document and here lies the first and major input problem. The fact that the card must be produced manually limits the speed of the system to the speed of the staff operating the card punch equipment regardless of the fact that cards can be read at the rate of 1000 to 2000 per minute . . . The average key punch operator works at a speed of 8000 strokes per hour . . . Verification cuts the production speed exactly in half and thus the net punched and verified card production is 50 cards per hour per card-punch machine." 143 Since a single card contains only the number of symbols or words that can be accom- modated in the space of 72-80 characters, it is ob- vious that preparation of even bibliographic cita- tions alone can be a slow and costly business. As noted in section 3.6.5 of this report, the tape typewriter is one of the devices that has been sug- gested for cooperative interlinkage between the document centers and between the centers and their customers at remote locations as in the "Reac- tive Typewriter" Plan (Mooers, 1960 [393]). Tape typewriters are also claimed by Horty and Walsh to be more economical of key-stroking time, such that "Flexowriter operators can produce be- tween 1400 and 1800 lines per day of statutory text," whereas "key punch operators used in pre- vious experiments could punch approximately 100 lines per hour of alphabetic materials, but could not maintain this rate for a sustained period of time." 144 Bernstein (1962 [57]) and Bernstein and Meyer- Uhlenried (1963 [58]) have made surveys of keyboard devices suitable for recording biblio- graphic data for documentation activities at EURATOM. They conclude: "For recording bibliographical data card punch- ing typewriters should be used if: the cards are directly fed into EDPM, the cards directly serve as storage and retrieval medium on conventional punched card machines, the initial document has no more importance than a proof reading sheet. Tape punching typewriters under program con- trol should be applied if: trained typists are available, repeated duplication with putting as heading of filing criteria is wanted, an easy tape- to-card-conversion is to be used. Tape punching typewriters without program control should be used if: the recording and correction process is to be kept as simple as possible, repeated duplica- tion without selection of filing criteria is wanted, the less expensive recording equipment is to be used." 145 In any case, intermediate transcription and conversion operations may be required in addi- 143 Cornelius, 1962 [1361, PP 44-45. 144 Horty and Walsh, 1963 T273], p. 259. 140 Bernstein and Meyer-Uhlenried, 1963 [58 J, pp. 75-76. tion to the initial keystroking and verification. A particular application may require a sequence of keyboard-to-punched-tape, tape-to-card, and card-to-magnetic-tape operations. Here also is a problem to be faced in the preparation of material in machine-usable form: the question of availa- bility and cost of suitable conversion equipment, and whether it can handle, for example, 5-, 6-, 7-, and 8-channel tape produced by different varieties of tape typewriters. Similarly, in the Flexowriter preparation of the texts of statutes at the Health Law Institute, Uni- versity of Pittsburgh, the records themselves might be of indefinite length, but "since the punched paper tape is being prepared for con- version to cards before being placed on tape, each flexowriter line corresponds to one punched card and is limited to 75 of the 80 possible characters. The last five spaces on each card are reserved for a machine code.'''' 146 This raises the general question of whether or not the machine-usable record or storage media is physically compatible with the machine processing equipment available. The most obvious example is that Remington-Rand punched cards, with 90 round hole columns which may or may not be punched, cannot be read with equipment designed for handling 80 rectangular hole columns on IBM cards, and vice versa. The possibility for com- patibility is nil. Convertibility might be achieved if suitable conversion equipment were available and if the economics of using such equipment were better than those of interpreting the information from one card set and rekeypunching it into the other type of cards. A more recent example, involving the general question of intra-Government and Government- to-non-Government cooperation and with regard to the problems of machine compatibility is that of the MEDLARS system. To facilitate use by other centers and libraries of materials in the MEDLARS files, it is planned to deposit copies of the tape files in various decentralized locations for easier access on a regional basis. Following a presentation of this system (Taine, 1964 [552]), R. T. Esterquest made the following remarks: "But there are more problems to a decentralized MEDLARS than meet the eye. In the first place, MEDLARS in Bethesda is centered in a Honey- well 800 computer. The computers locally avail- able to my library happen not to be Honeywell machines and cannot read Honeywell's tape." 14T To which the speaker replied : ". . . Honeywell is not alone among the computer manufacturers in being incompatible with the other computers. In fact, there is even a considerable degree of incom- patibility between different computers from the same company. However, this is not a serious obstacle to the effective distribution of duplicate MEDLARS tapes to outside computer centers. ""Hortv and Walsh, 1963 [273], p. 259. "'Taine, 1964 [552], p. 130. 115 The conversion of M-H 800 magnetic tapes to other tape formats is economically and technically feasible at the present time." 148 A case in point is the Biomathematics Division at Ft. Detrick, Maryland, which has the Honey- well tapes converted to its own UNIVAC Sys- tem tapes, via an intermediate conversion to IBM tapes, in order to incorporate MEDLARS material into its own search files. But, as Esterquest fur- ther remarked, "even if Honeywell's tapes were to be converted, there is still noncompatibility at the systems or operational level." 149 The possibilities for machine convertibility be- tween systems have been little explored. The few exceptions have mostly been limited to mechanical transcriptions and transliterations, resortings and reassemblies, often accomplished by slow and cumbersome auxiliary equipment. For example, although a UNIVAC tape version of 400,000 words of Webster's 2d International Dictionary was available to RCA researchers, it was not usable with any of the computers to which they had access. Accordingly, it was necessary to reprocess the hundreds of thousands of the original punched cards to produce another magnetic tape accept- able to the 301. More important from the standpoint of future compatibility and interchangeability of machine- usable products, however, are those problems which arise with respect to "format control, questions with respect to the conventions to be used in key- punching, and problems of preediting." 150 These problems in turn are closely related to the possi- bilities of standardization of codes and conven- tions for recording, keypunching, transliteration, and transcription. The variety of input characters where text — even if only text of authors' names, titles, and other elements of the typical descriptive cataloging record is involved — adds significantly to time and cost of preparation. Thus manual editing, trans- literation, encoding and simplification before in- put is typically required, in accordance with rules and conventions such as for the spelling out of chemical symbols or Greek letters and for consist- ent practice with respect to format indicators and font change designators. Nugent (1959 [431]) and Ray (1962 [471]) have pointed out some of the requirements and some of the difficulties in providing a manual for keypunching of natural language texts. New- man, Swanson and Knowlton provide (1960 [428] ) an even more complicated notation-transliteration scheme necessary for keypunching of patent dis- closures in which a wide variety of boldface, italic and special symbol inserts have been accounted for. Patrick and Black have recently suggested that "in the following rather simple way, a catalog card could be easily punched : start with a mode- change character that set [s] the mode to bold-face 148 Ibid, p. 131. ""Taine, 1964 [552], remarks of R. T. Esterquest. "° Stevens, 1962 [543], p. 63-64. alphabetic ; following this would come the author's name; following the trailing punctuation would be a mode-change character that specified the mode as numeric ; the author's year of birth would fol- low ; whenever this line was complete, an end-of - the-line symbol would establish the format for the title line; and so on down the card, with the changes of font, capitalization, intervening punc- tuation, through the end of the recorded infor- mation. If the card were multilingual ... a mode-change character would indicate this fact. . . ," 151 Font designators available in one experimental program for the "extraction of significant infor- mation items to be used in the automatic produc- tion of secondary publications from a formatted, punched tape manuscript" include signals for changes from Roman to Greek alphabets, subscript or superscript, italics, bold, sans serif. 152 We note, however, that the number and identity of these mode-change characters must be estab- lished and their usage for a particular applica- tion must be defined. This requirement raises the next specific problem, that of character sets for input, internal processing, and output. 4.3.4. Character Sets for Machine Processing The present disparity between the very large character sets which typically can appear on the printed page or even on the catalog card and the size of the character sets available on typical type- writer, keypunch and console keyboards creates a dilemma. Either much of the information con- veyed by typestyle changes, boldface, superscripts and subscripts, special symbols and the like must be eliminated as part of the input process or it may be preserved by spelling out, encoding, and the use of special clue symbols, sometimes amount- ing to three or four for each "text" character. If the information is eliminated, then its restoration is generally not feasible, so that the output char- acter set will usually be limited to not more than 64 letters, numbers and symbols, printed in upper case. Yet what is desired is "legibility to permit the rapid scanning of the printed product [which] is an important requirement that implies a capa- bility for printing in a variety of type fonts and sizes .... A variety of type styles, weights, and sizes are needed to meet the standards of Library publication." 153 The present alternative, however, pending the development of automatic, multifont, page reading devices, is likely to increase the already stag- gering costs of input data preparation several- fold. "Obviously, direct conversion of data by machine scanning would be infinitely superior." 154 Unfortunately, the third possible method of input of bibliographic and textual information men- tioned earlier, that of the use of automatic char- 151 Patrick and Black, 1964 [450], p. 41. ,M Lundy, 1963 [361]-, p. 115. 1R *King, 1963 [318], p. 9. 1M U.S. Senate, 1960 [592], p. 38. 116 acter recognition techniques, is not yet feasible. In general, a summary appraisal of the state of the character reading art made in 1961 is still applicable: "With character recognition equip- ment, the printed or typed page would be inserted into the reader, which would scan each line, con- verting each letter recognized into the appropriate machine code language equivalent. Reader out- put might be punched cards, punched tape, mag- netic tape, or directly to computer. Prototype page readers for good quality printed text in one or several different type fonts are already under- going active development and testing. . . . "The feasibility of machine reading of good quality printed, typed or, in some cases, even hand- printed characters, has been successfully demon- strated. However, there is little operational experience to date except with limited vocabularies of specially designed characters and questions of equipment cost and confidence in performance have so far precluded general use of readers for preparing texts in machine-usable form. In addi- tion, if the source material contains graphic material, equations, and formulas as well as text, it may be necessary to perform extensive manual editing and masking out before input to a reader." 155 More recent developments involving design pro- posals and demonstrations of page-readers with some multifont capability are still far below the character set size required for most scientific and technical literature. Shaw reports, for example: "To get an indication of the number of difficult alphabetical characters that may be used in a single scholarly book, I asked Mr. Theodore Bes- terman to count the different pieces of type used in his World Bibliography of Bibliographies. This he did, and as shown in the preface to the third edition of this work, he used approximately 1950 different pieces of monotype . . . There are a number of alphabets that are not included in Best- erman, as well as mathematical symbols, etc." 156 Even for the case of catalog cards only, the severe problems of machine-reading are those of mixed-font, mixed-weight and size, mixed alpha- bets and formats, and varying conventions of abbreviation, punctuation, and representation which have been adopted to differentiate the var- ious types of selection, locator or reference clues in a manner that would be convenient for the hu- man eye in scanning and filing. This is why it is only reasonable to conclude as of today that" . . . The automatic conversion of files, such as the Na- tional Union Catalog, which contain considerable heterogeneous symbolism, is not likely in the near future." 157 The question of how large input and output character vocabularies should be is further illus- trative of the ramifications, in a specific problem sense, of the general problem areas, including '•■• Stevens, 1962 [543]. pp. 61, 63. i 6 " Shaw, 1962 [518], p. 268. 1BJ King, 1963 T318], p. 20. those of problem definition (especially, of "seeing the problem and seeing it whole"), languages, and user acceptance. Clapp has spoken to the prob- lem of the so-called "upper-case limitation" as follows : "As long as Mr. Watson would only give us a character font in capital letters there could not be any great fervor to load. If I may say so, the whole picture of automation in libraries from the thirties right down to the present date has been controlled by that uppercase limitation." 158 Regardless of the size and contents of input and output character sets, however, there are additonal special problems of the internal character set which will affect planning for sorting, ordering, filing, and interfiling. "Even though we have not yet agreed on a standard character set for our data processing computers, each computer has a single set built into it. This set is the 'natural' set of that computer, and all other sets are defined in relation to this built-in character set. The character set is extremely important in file definition since the character set defines the order of a file once it is sorted." 159 An example involving computer maintenance of a thesaurus in the Public Health Service uses machine-generated line numbers for each diction- ary line because "filing of terms cannot be done in a strictly alphabetical manner. For example, X-RAY is filed under X, while D-GLUCITOL is filed under G." 160 The effect of the particular internal set in a particular computer on filing and ordering also points up the importance of either exact compati- bility in input material exchanged between organi- zations or the availability of carefully program- med conversion capabilities. For example, in certain computers, sequences of information recorded in punched paper tape must be interrupt- ed at arbitrarily fixed intervals by a special symbol such as the carriage return in order to prevent shifting in input so that one or more bits are de- leted or added to the original bit pattern. If in the tapes received from another installation there is any deviation in position or bit pattern of the interrupt indicia, such shifting may occur in any case. This will result in garbling of the encoded text. Thus, for example, a one-bit shift of the text of the title : "Real-Time Data," with upper and lower case as recorded in ASCII code might result in the following message being actually read by the machine: ")206 (SYNC) : 462 . . ." If and only if such shifts were consistently made for all inputs and a compensating shift made on output, could such messages be used for file posting, file main- tenance, or retrieval purposes. 4.3.5. Internal Processing and Programming Languages Specific questions of internal processing lan- guage and programming language are interrelated 158 Clapp. 1964 [114], p. 54-55. 150 Patrick and Black. 1964 [4501. p. 31. 100 McGee et al., 196.3 [377], p. 348. 11^ with those of internal character set, fixed versus variable word length allowed in internal memory, and compatibility, or interchangeability. For the 21 different information retrieval languages and systems listed by Grems in 1961 ([227]), 11 dif- ferent computers are used. Some use a binary internal language, some a binary-coded decimal, some a full binary-coded-alphanumeric. Special programming languages are used, for example, in BASEBALL (an experimental question-answer- ing system), for the ACSI-MATTC intelligence processing system, and for linguistic data processing programs at the System Development Corporation. A computer program to produce printed book catalogs at Boeing Research Laboratories, for another example, has the following features which illustrates some of the problems of compatibility : (1) it is written in FAP language for an IBM 7094-1401 intsallation but because it uses only three index registers, it can be "directly adapted to IBM 7090 or 709 with only minor revision" ; (2) it provides subject entries, conventional title en- tries and permuted title entries, but the latter are based on manually preselected key words and "a nonprinting symbol entered on input copy causes rotation of the title about the normal indexing locus, the left margin, with no loss of context, whatever its length" ; (3) input can be either from the preferred form, Flexowriter tape, or from standard punched cards; (4) "sorting is accom- plished in a manner analogous to Library of Con- gress rules, as commas and dashes in subject head- ings are used as control characters to permit the necessarily unique subject-title intermix." 161 "Program revision" however minor, "nonprinting symbols," "control characters," "normal indexing locus" are all clues to difficulties of interchange- ability even with organizations having a closely similar equipment configuration. In general, with respect to computerized docu- mentation operations little or none of the input, intermediary, or output products are directly usable in any other system whatsoever. Apparent exceptions, such as the distribution through the SHARE system of various K7WTC index prepara- tion programs, require innumerable compromises between what-is-desired and what-can-be- obtained both cheaply and expeditiously. Even for the case where the standard Luhn KWTC format is only slightly modified, input copy is generally not interchangeable today. Two additional examples of related problems might be mentioned here. The first has to do with the wide variations in length and contents of the stop lists used to delete from indexing considera- tion the common and insignificant words in KWIC and similar type machine-prepared indexes. The length and contents of stop lists used directly affect the number of pages com- piled, and hence both the costs and the usability »« Weinstein and Spry, 1963 [638], p. 233. of the printed indexes produced. Changes in stop lists over time in a single organization, which is quite common practice, will seriously impair the prospectives for cumulative indexes of this type. The second example also relates to computer production of permuted entry indexes, but in this case subject indexes involving a classification scheme. Mills says : "The wide use of the UDC in published indexes and bibliographies as well as by journals and abstracting sendees makes con- sistency between these in the use of UDC desirable. To assist this, some users have advocated the ex- tensive use of the colon device in place of special (and even common) auxiliaries in order to facili- tate the routine production of permuted entries, e.g., 669.295 : 669.018.5 . . . would be preferred to 669.295.018.5 . . . There is no general agree- ment on this however." 162 For another example, Wofsey in his analysis of four COBOL compilers for four different com- puters (1962 [651] ) found an average of more than 21 variations which would necessitate changes if a source program from one machine were pro- cessed on another. These differences involved compiler restrictions in one system but not the other, options available in one and not the other, failure to accommodate changes in COBOL speci- fications (1960 as versus 1961) and variations in the equipment. Further, at a seminar on the prospects for an information retrieval programming language, the chairman summarized the consensus as follows: "The wide divergency of basic file layouts, index- ing techniques, etc., make the development of a 'standard' IR language impractical at this time. Thus, the answer ... is to encourage the develop- ment of special-purpose IR languages to meet individual needs, even though this complicates the translation problems enormously . » 163 5. Implications for Further Progress Observations that may be made as a result of this literature review are as follows: certainly in the United States, cooperation has long been evident among librarians since the establishment of the American Library Association in 1876. Interna- tional aspects of such cooperation are evidenced by such widely separated events as the 1908 Anglo- American rules for cataloging author and title entries and the 1961 International Conference on Cataloging Principles, sponsored by the Interna- tional Federation of Library Associations. Cooperation is also evident among those active in the development of new and less conventional pro- cedures and equipment, especially since World War II. Examples of this interest may be seen in the development of a standard for the descriptive cata- loging of government reports under the sponsor- ship of the present Committee on Scientific and 162 Mills, 1964 [388], p. 47. KBSammet, 1962 [503], p. 8. 118 Technical Information, in the NASA, AEC, and DOD microfiche agreement and in the approval by the American Standards Association of an American Standard Code for Information Inter- change (ASCII). Most of these examples of cooperation are con- cerned either with physical aspects of the exchange of information or with the identification and de- scription of documents. Cooperation in the "tag- ging" of the subject content has not been so prev- alent. The COSATI standard, for example, makes compatibility possible among several agen- cies in describing bibliographic items. The achievement of compatibility among three Fed- eral agencies (AEC, DDC, and NASA) for proc- essing these reports has been difficult to attain, even the achievement of convertibility among the terms used for indexing content proving elusive despite investigations such as the Datatrol studies which look toward a scheme of generic subject cate- gories that will subsume the composite subject con- tent of the research reports dealt with by these agencies. Major deterring factors in this area of potential compatibility or convertibility appear to be those of achieving consistency of human indexing and of bridging the unfortunate gap which still ap- pears to exist between librarians and "information scientists," between traditional libraries and infor- mation centers dealing primarily with the report literature. While the original impetus for this review was with respect to the implications that might be found regarding machine language compatibility, it appears that very little helpful precedent is available. The history of both early and current activities point to a serious question of prematurity of compatibility efforts in this area, much less ef- forts directed toward standardization. It would appear that firm standards for report numbering compatible with an identification system for the journal literature, for descriptive cataloging em- bracing both the traditional and the more recent types of publication and disclosure, and for subject cataloging and abstracting at least to the extent of minimum, classificatory schemes or indexing vocabularies, are all prerequisite to efforts in the machine language area other than those of agree- ing upon such detail matters as the common coding of a minimum character set (ASCII) . "For the past ten or so years, we have been pre- occupied with the development of new techniques, few, if any, of which have yet proven themselves to the exclusion of others. Now, with small chance of consensus as to the best approach for a single situation, talk of compatibility emerges with stand- ardization looming as its ultimate manifestation. Premature standardization on a large scale to sys- tems which are not the best is entirely possible." 1 Y. S. Touloukain stated at Congressional hear- ing in 1963 that "one cannot be against standardi- 1 Anderson, 1962 [29 ]i, p. 116. zation, the question is what it is that we want to standardize. . . . Industrial compatibility in nuts and bolts and machine threads is one thing, and compatibility of coding intellectual information which may be in numerical form, pictorial form, quite often in the form of opinion and judgment, is another." 2 The problems and difficulties include, as we have attempted to illustrate, those of defining require- ments throughout the entire cycle of information processing from generation of basic data to use of recorded knowledge, those of language at many levels, those of human consistency and human ac- ceptance. Problems raised by prospects of mech- anization include: (1) the physical character- istics of storage and recording media, (2) layout and format of storage and recording media, (3) selectivity as to units-of -information to be handled and items to be recorded, (4) layout and format of identificatory indicia, (5) coding required and conventions used for abbreviations, variant spell- ings, and transliterations, (6) code symbol vocabu- laries, (7) machine language variations, and (8) programming language variations. Each of these problems obviously has impact on the underlying intellectual, organizational, and systems problems, reaching back into traditional library practice and conventional interlibrary co- operative activities. In the 1961 report on Senate hearings with respect to the coordination of scien- tific information, continuing and intensified efforts were recommended in such areas as " ( 1 ) Formulat- ing a thesaurus of index titles for classifying re- search and development on a Government-wide and national basis; (2) Developing standardiza- tion and compatibility of data processing systems, at least for all Government agencies ; (3) Correlat- ing a central project registry with the indexes maintained by different agencies, with informa- tion and data centers, with research documenta- tion and with the NSF Register of Scientific and Technical Personnel." 3 These objectives remain worthy long-range goals, but there are many practical difficulties which must be faced. Thus, Wall in commenting upon the implications raised by the above three points suggests : "Specifically, a thesaurus for government wide application would seem to be useful as an inter- agency communication vocabulary standing for concepts of interagency interest . . . but each agency would need its own thesaurus, consistent with the interagency one, to cover its own unique terminology plus that of the interagency thesaurus which may be of interest to the individual agency. An interagency organization thesaurus such as this for government-wide and national application is out of the question, I believe, until government and non-government organizations involved have first created their own thesauri, thus creating a base J U.S. House, 1963 [582], p. 191. 'U.S. Senate, 1961 [593], p. 4. 119 for common agreement and resolution of individ- ual disagreements. "Standardization of all government data proc- essing systems might well turn out to be a will-of- the-wisp — at least for a number of years. There are such tremendous problems involved, especially considering all the levels at which standardization may be striven for. At what levels is standardiza- tion to be achieved and with what economic justi- fication (bit structure, collating sequence, charac- ter sets, word length, programming language, rec- ord length, record format, media, etc., etc.) ? "Finally, as for 'correlation' of project data and information, we can see the possibility of SIE (or equivalent) and similar agency indexes being coor- dinated, but it seems to us that beyond that the 'implication balloons to a gargantuan task — namely, correlation with information centers' and with 'research documentation. . . .' " 4 Under these circumstances, as Fussier said a decade or so ago, "the conditions for successful co- operation are both limited and important." 5 It should, of course, be noted that many of the recommendations of the Crawford and Weinberg panels, of Congressional committees and subcom- mittees, of the Science Information Advisory Council, FASCI, COSATI, the DOD-wide docu- mentation coordination program, the Clearing- house for Federal Scientific and Technical Infor- mation and others which have been put into effect will have a continuing impact on increased co- operation, compatibility, and convertibility. The COSATI task groups on technical vocabulary compatibility, on conspicuity, on descriptive, cataloging standards, on a national inventory of journal literature, on a single Government de- pository system and on research and development in the information sciences all point in helpful directions. Nevertheless, a variety of questions and pro- vocative suggestions are implied by the reports in the literature of prior experience, failures and shortcomings, the lack of extensive examples and the recognition of new challenges and new op- portunities. First and foremost of the implica- tions is that: ". . . While the theoretical recog- nition of the advantages for co-operation can be taken for granted, it is a much more difficult thing to achieve real and helpful co-operation . . . ." 6 Honest differences of opinion have existed and continue to exist with regard to centralization and decentralization, mission versus disciplinary orientation, Governmental coordination and pri- vate initiative, prematurity or belatedness of standardization, precoordination or postcoordina- tion in indexing, down to details of physical char- acteristics of storage media and the like. For example, Agard Evans has noted that despite an international agreement for cooperative abstract- ing in building documentation, only nine of the 15 participating countries conformed to the agreed- upon standards, even in such details as physical characteristics of products. He reports, for ex- ample, that in France "the abstracts are produced in the prescribed form but printed on thin paper and not on cards." 7 Nevertheless, "if informa- tion systems are to succeed they must be uniform enough to serve the common needs of various cate- gories of users yet flexible enough to contribute to specialized needs." 8 Compliance with standards, or agreed-upon conventions of practice, is also difficult to achieve. For example, the Unesco recommendation, "Guide for the preparation and publication of Synopses" has been "officially adopted by the International Council of Scientific Unions but seems to have had little publicity judging by its slowness to take effect." 9 As another example, Cady (1950 [92]) reports on the survey of 40 general periodicals and 25 scientific journals with respect to their compliance with "An American Standard Reference Data and Arrangement of Periodicals, Z 39.1-1943", which was intended to prescribe a standard location in journals for the date, volume issue, page numbers, table of contents, authors and titles of pages. For the general publications, it appeared evident either that the editors had never heard of the standard or were ignoring it, and, while the technical journals adhered in some cases, considerable im- provement was indicated. Provocative suggestions have been made with respect to a variety of new and specialized types of clearinghouse, directory, and exchange services, and for the extension of existing services. Guides or directories to sources of information have been advocated at least as early as the Royal Society Conference on Scientific Information in 1948. 10 Directories of specialized information centers and services for both the physical and biological sciences (National Science Foundation 1961 [419]) and of the information resources of the United States (Library of Congress, 1964 [339]) are examples of recent approaches to one aspect of this problem. However, although "we have in- formation about information centers in the United States ... we do not have adequate information about centers in other countries that might well provide service to us." u Closely related, as anticipating new services for continuing systems design, systems evaluation, and systems improvement is Bourne's suggestion for a "documentation census": "One approach that would be very useful would be to establish a permanent mechanism for obtaining current and accurate information about the parameters of the literature problem — that is, a continuing docu- 4 Eugene Wall, private communication, March 30, 1965. 5 Fussier, 1953 [208], p. 222. a Readett, 1960 [472], p. 155. 7 Agard Evans. 1959 TR], p. 493. 8 H. H. Humphrey, "Findings by Subcommittee Chairman," in U.S. Senate, 1961 [590], p. XVI. 9 Ci-aig. 1964 [146],. pp. 4-14. "Urquhart, 1952 [604], p. 236. "Heumann, 1962 [258], p. 122. 120 mentation census that keeps a finger on the pulse of the literature flow to note the volume, charac- ter, and trends of the various mixes hi the flow . . . ," 12 Another suggestion is for the provision of photo- copies for all items covered by a given abstracting or indexing service, or for an entire broad field of science. Wilkinson and Waldo ask : "Couldn't someone assume the responsibility to provide photocopies of any article which has been ab- stracted in C.A.f .... What we really need is a sort of clearing house to which we could send orders for copies of any articles which have ap- peared in the chemical literature. The clearing house would be able to provide copies immediately from its own holdings or forward the order to a source which they know can definitely provide it immediately." 13 Taylor, writing in 1963 with respect to increased cooperative action among Federal libraries, sug- gested that among the particularly timely joint measures that might be undertaken would be the establishment of a clearinghouse on bibliographi- cal projects and a review of translation services: "demand, cost, speed, method of reproduction, and indexing." 14 Also, with respect to technical translations, the general suggestions find echo in a specific case of mechanization: "One of the noteworthy features of the French system is the recent introduction of a Selecto-Cordonnier mechanical selector, which enables translations to be traced even from incom- plete data such as only the author's name, subject of the article, title, or date of periodical from which the article was taken." 15 In the area of translations, it has also been sug- gested that directories or records of holders of copyright should be established, especially with re- spect to obtaining authorizations for translation to be made. Reed says, further : "I think we should explore the possibility of more frequent recourse to owners. ... It might be well worth while con- sidering the establishment of some focal point at which experience of owner approach could be ac- cumulated and advice given." 16 A more general approach to the copyright owner clearinghouse has been reported. "A grant has been made by the Council on Library Resour- ces for a feasibility study of a clearinghouse which would serve as an intermediary between owners of copyrighted material and scientists and others wishing to make one or more photocopies of it. The study is being conducted by the Committee to Investigate Copyright Problems Affecting Com- munication in Science and Education." 17 Another example of a specific suggestion or question is whether a "Current Contents" project "Bourne, 19G2 [73]'. p. 162. " Wilkinson and Waldo, 1962 [646L, p. 175. Note that VINITI does provide such services, see p. 49 of this report. 14 Taylor, 1963 T565], p. 77. "Liebesny, 1960 T340], p. 150. M Reed, 1960 T475], p. 166. « Scientific Information Notes 5, 4, 12 (1963). can be established for multisubject technical re- ports, especially consolidated progress reports of major research laboratories I The idea of a central clearinghouse for "field-of- interest" registers extends itself quite logically into the idea of one or more relatively centralized and specialized selective dissemination operations. New machine possibilities for the production of special-purpose and 'tailored' bibliographies on demand raise the further intriguing prospect of demand specified in advance, of "profile bibliog- raphies'' as a further extension and development of selective dissemination. It has recently been recommended, for example, that pilot trials be conducted for "screening com- puter tapes of current references to bio-medical literature, such as those produced in the MED- LARS program, to provide individual bio-medical scientists with a current awareness service spe- cially tailored to their interests, habits and preferences." 18 Fundamental questions that are raised, however, go to matters of bridging gaps between libraries and report handling centers, of maintaining prop- er balance between requirements of large-scale or centralized services and local needs, of developing systems with sufficient flexibility and provision for the incorporation of user feedback and usage data. In the area of U.S. Government responsibilities, for example, has enough attention been given to departmental and agency libraries — their policies, procedures and problems? The Brookings survey suggests not : ". . . It is fair to say that, taken as a group, the libraries of executive departments and agencies have received little concentrated at- tention either from government policy-making officials or from students of government. No gen- eral policy regarding their functions has been enunciated ; no standing body of administrators or librarians is concerned with their problems; and no current and comprehensive statistics have been available on the magnitude of their holdings, the cost of their operations, or the range of their services." 19 What measures of compatibility can be assured in planning or implementing improved and in- creased Government-wide services in terms of accessibility and costs of suitable equipment for the use of mechanized records, microforms, and other products available only hi machine-usable form? Priorities of access to centralized serv- ices? Compatibility or at least convertibility of physical form and medium, content, and format of bibliographic control information with special- purpose needs and local secondary handling costs? The latter may include, for example, the human costs involved in scanning indexes, abstracts, and announcement bulletins which are too lengthy, too cumbersome and which have poor readability because of "upper-case limitation." 18 "Communication Problems In Biomedical Research" 1963 [403], p. 21. "Orlans, 1963 [434], p. v. 121 In this connection, the specific question of the availability of inexpensive microform readers also arises. As Becker points out : "Little information on machines and other equipment was collected in Dr. Evans' survey. In their questionnaire replies a few librarians did report the presence of micro- film reading machines and photoduplicating equipment, but the data were too sketchy to permit significant conclusions. Conspicuous by its absence in their comments, however, is any widespread con- cern about equipment and the impact which au- tomation is apt to have on their libraries. If the federal library system is to be strengthened this certainly is one area that needs detailed attention and helpful support." 20 Turning, then, to the more specific area of present and future prospects for mechanization, we find suggestions for the provision of basic tools such as the National Union Catalog in "machine- language form, periodically updated and main- tained," 21 for "a systematic program of training and education in documentation [which] should be developed by federal libraries to encourage the automation of their activities, whenever practi- cal," 22 and for the encouragement and support of cooperative experimentation in mechanized documentation. Among the improved clearinghouse and ex- change services which might be set up for the me- chanization area would be exchanges of indexing and cataloging information tapes, exchanges of computer programs for listings and selective as- semblages of such information, exchanges of pro- grams for conversion of one machine language to another or to provide magnetic tape images of punched paper tapes so that information in various formats can be "unpacked" and reformatted by computer, and exchanges of machine-usable texts, records, stop lists, word lists, dictionaries and thesauri. We note, however, in this last example, some warnings of possible prematurity. For example, in House hearings on research in mechanized trans- lation the following exchange occurred: (The Chairman) "Each group doing research in this field builds its own dictionary. Why can't the research work on dictionaries be exchanged?" (Mr. Dostert.) "We should perhaps make a dis- tinction between 'word lists' and 'dictionaries.' There seems to be no complication in providing for the reciprocal exchange of word lists developed by the various groups. It seems to me also that some coordination and standardization could be arrived at in regard to keypunching procedures and format. However, if by 'dictionary' we mean the lexical materials plus the codes for the analyti- cal operations, this would seem less feasible, since the codes are devised on the basis of the particular philosophy and techniques pursued by the several groups." 23 Similarly, King says: "It is too dangerous to insist upon or freeze on a 'programming language' for lexical material, but some sort of clearinghouse will soon be essential in view of the manpower requirements." 24 Another aspect of possible prematurity with re- spect to mechanized systems relates to the more general question : Should a uniform system for at least minimal subject content indication be based upon a universal classification system, facetted classification schedules, alphabetized subject head- ings, thesauri of descriptors, Uniterms with scope notes, "free indexing" (largely of words derived from text) , or mechanical derivative indexing by automatic extraction of keywords or words occur- ring with significant frequencies in a text? An example of desirable research in this area might be the project in which "AEC . . . cooperated with the National Science Foundation in the develop- ment of an experimental classification and coding system for AEC documents. The ultimate purpose is to develop a coding scheme which can be utilized by a small digital computer which would be eco- nomically feasible for smaller libraries." 25 Further, is the code or notation scheme by which the classification or indexing terms are recorded sufficiently flexible for multipurpose use? It should be remembered that the major criteria for notation schemes are "usually judged to be hospitality (the ability to incorporate new classes in their correct, logical place . . .) , simplicity (the ease with which symbols convey position) and brevity (a central element in simplicity) ." 26 Here the problem of possible prematurity relates the evidence, on the one hand, of the serious lack of consistency between human indexers, even when they are using the same controlled indexing or classificatory vocabulary, to that of the, at the least, provocative results from automatic assign- ment indexing and automatic classification experi- ments, on the other. While in the latter case the experimental data available to date have been limited to very small test samples, to a very small number of subject headings or indexing terms to be assigned and to quite carefully selected pseudo- collections within highly specialized fields, within these limitations the results do suggest a level of indexing such that users, or indexers, would agree as to the relevance of terms assigned by machine about as frequently as they now agree with each other. 27 This is at about the 50 percent level, or less, for routine mass-production indexing opera- tions, although it may rise considerably higher for a carefully trained, highly motivated staff as at 20 Becker, 1963 [46], p. 92. 21 Swanson, 1964 T549], p. 87. 22 Becker, 1963 [46], p. 92. 23 U.S. House. 1960 [588], p. 107. 2 * King 1964 [319], p. 241. MILS. Senate. 1960 [592], p. 37. 26 Mills, 1964 [38S], p. 36. 27 For a review of some of these experiments see Schultz, 1964 T510]. A state-of-the-art review of automatic indexing develop- ments is also in process of publication as an NBS monograph. 122 Chemical Abstracts or among those participating in the Cranfield experiments. It is clear that any successful interchange of compatible or convertible materials with respect to subject cataloging or content analysis must be dependent upon a very high level of intrasystem consistency as is, of course, optimal use of the individual systems. To date there is hardly a sys- tem that has approached the degree of consistency which will provide such optimal use of its own facilities or of the export of its products to other systems. Further, as Painter points out (1963 [445] ), convertibility between indexing systems is highly dependent upon a high degree of con- sistency within each system. Perhaps only by mechanized processes of indexing and classifica- tion can results be replicated and, hence, con- sistency achieved. This area, therefore, certainly appears to be one in which future research should be encouraged and supported. "Research should be expanded as to how far individual systems — manual or mechanized — may vary one from the other (as they do vary at pres- ent) while still offering reasonable opportunity for compatibility." 28 Other research areas in which greater coopera- tion, coordination and encouragement would appear indicated are those of systematic attack on the bottleneck problems of input and of the de- velopment of remote consoles and other means for man-machine interaction in the information system. The first requires further research and development in the field of automatic character recognition, with emphasis on multifont page- readers and other alphabets than the Roman. As noted previously, the cost of converting the present union catalog information of the Library of Congress, assuming keystroking of the informa- tion for each entry, would be not less than $7,500,- 000, and in the case of Federal Departmental Libraries the sheer costs of transcribing to punched cards merely the descriptive cataloguing information for these holdings would exceed, at minimum, $10,000,000. Pending development in character recognition, therefore, it has also been suggested that the possibilities for machine proc- essing of stenotyped copy should be further explored. 29 A somewhat similar question of adequate character vocabularies also arises in the research area of man-machine interface. t "Machine methods of editing, spelling correc- tion, type selection, page formatting, and line justification have all been demonstarted. Stress must be placed on software techniques for editing and page composition. Computer graphic com- posing equipments must be made capable of pro- viding high quality, reproducible copy at rates of several hundred characters per second with a large repertory of symbols, characters, and type sizes. . . . There should also be some means, whether at the console or central processor, of automatic format control and a means of auto- matic transliteration of words in digital storage to the Roman alphabet, regardless of the source language alphabet." 30 The King report also emphasizes some of the benefits to be gained by the development of adequate user consoles as follows: "users' annotations of subject classification or assignment of subject headings could be accepted as input at the console, subjected to further review and edit- ing by a librarian and then incorporated, as desir- able, in the system. Users could also comment on the similarity or relatedness of specific papers, reports, or books which they utilize." 31 "A capability for browsing by use of the console should be of even greater significance. The op- portunity to examine statistical data on the num- ber of entries in a bibliography, the bibliography itself, and then selected pages of particular items which may include title pages, tables of content, and indexes, all on a successive rapid response basis and coupled with subject access to whatever depth economics permits, will provide a far more flexible intellectual interaction between the user and the collection than occurs in wandering through the stacks as though one were shopping in a supermarket." 32 A final further example of needed research is in the area of programs and programming languages for index compilation, linguistic data processing and automatic content analysis. Is it feasible to develop a general purpose compiler to cover a variety of output formatting requirements, includ- ing those for various automatic photocomposition systems? Is it possible to develop a problem- oriented interpretative language to do for infor- mation selection and retrieval wjiat ALGOL and COBOL offer other types of computer applica- tions ? Can Federal support as well as encourage- ment be provided to those who are capable of determining answers to these questions? "The federal library community offers an ideal setting for the establishment of a cooperative auto- mation program involving education, research, and experimentation. No effective means exists today for federal librarians to exchange and evaluate their knowledge and experience of mech- anization, or to pursue broad, government-wide mechanization programs. . . . The value of co- operative experimentation in library automation should be emphasized in order to minimize cost, avoid duplication, and ensure widespread dissem- ination or results." 83 However, notwithstanding the probable results of further research, the problems of the handling of scientific and technical information are urgent ^Cahn, 1962 [04], p. 27. "King, 1963 [318], p. 0; Patrick and Black, p. 43. 1964 [450], 80 King, 1963 [318], p. 20. 31 Ibid., p. 11. 82 Ibid, p. 23. 83 Becker, 1963 [46], p. 93. 123 today. While information handling of the scien- tific and technical literature was principally a matter of manual typing, scanning, labelling, re- cording, storing and searching, the questions of compatability and convertibility arose only as a matter of avoidance of duplication of effort. Co- operation was and is still predominantly con- cerned with the sharing of information about hold- ings, availabilities, exchange of hard copy items and records such as catalog cards. The efforts are collaborative actions of libraries and information centers (as organizations) and of librarians and document alists (as individuals or as members of professional groups). With the introduction of machine techniques, however, cooperation in the conventional sense ac- quires one or both of two prerequisites — those of compatability and/or convertibility. What, then, are some of the practical steps which can be taken today, without unnecessary prejudice to future developments with respect to the intellectual organization of subject content analysis, automatic content analysis, machine search and retrieval, man-machine interaction dur- ing search and selection processes, and interlock- ing communication networks ? First and foremost, as we have seen, is to work for mutually acceptable conventions and standards with respect to journal abbreviations, translitera- tions of proper names, ordering and coding of corporate authors, form, order, and content of en- tries in biblographic citations. From the point of view of potential mechaniza- tion and of interchange of programs for, and prod- ucts, of machine compilations can agreement be reached with respect to the number of characters per line to be reserved for identifactory purposes so as to minimize, on the one hand, "double look- ups," and yet not to jeopardize contradictory re- quirements for multipurpose identifiers? Will it be possible to arrive at a Government- wide report numbering system? If so, can it be such as to incorporate security classification infor- mation ? Can it include an error-detecting or er- ror-correcting digit ? At the very minimum, there should be concern for machine transliteration pos- sibilities from one coding, cataloging, or indexing system into another. More significantly, every effort should be made to provide systems interface byproducts, such as minimum version catalog cards, including subject cards, that will be adaptable to less sophisticated equipment in smaller organizations and even to manual filing as suggested by Altman (1963 [12] ). The latter point raises certain further possibil- ities with respect to which an organization such as COSATI might provide leadership and initiative, at least for Government, by virtue of its influence and position. The first is to encourage the use of Federal procurement authority in such way as to promote greater compatibility in the equipment produced by various manufacturers and to insist, for example, not only for at least one input and output facility for handling ASCII and for punched card input and output capabilities as well. Somewhat less obvious, but equally pertinent to the elimination and reduction of barriers to scientific communication, would involve promo- tion of policy leading to more favorable copyright conditions, to push the "preparation, elaboration, and revision of classification systems in neglected fields," 34 and to press for changes in editorial, page makeup, and even typographical practice that will make typed and printed text more amenable to reading and processing by machine. A responsible, far-seeing coordinating body (even if its functions are advisory only) should, it is suggested, direct attention to problems and pos- sible solutions in the area of present barriers to scientific communication. Elimination or reduction of major barriers to more effective communication and use of scientific and technical information will first of all broaden the bases for cooperation as such. Second, such elimination or reduction of the communication barriers contributes directly to improved problem definition, determination of requirements, system design, and practical implementation. Finally, without such reduction, compatibility and/or con- vertibility within and between isolated subsystems cannot improve the overall situation and may in fact make it worse. We are especially indebted to those members of the Ad Hoc Committee who have forwarded com- ments and criticisms, many of which have been reflected in revisions to the earlier draft of this publication. In particular, we are indebted to Scott Adams, Deputy Director of the National Library of Medicine, and Ann F. Painter, and other reviewers not only for detailed and cogent criticisms but also for access to private files or per- sonal recollections. The cooperation and assistance of Josephine L. Walkowicz and her staff in checking the bibli- ographic references and of Betty Anderson, Cath- erine Bailey, Helen Grantham, Norma Lupowitz, and Anna K. Smilow in the preparation and proofreading of the manuscript are also gratefully acknowledged. Appendix. List of References Cited and Selected Bibliography [1] Abelson, P. H., Information transfer and retrieval, Science 141, 319 (1963). [2] Adams, S., Discipline- and Mission-Oriented Science in Relation to the Abstracting Problem, in R. A. Jensen [ed]. Proceedings of Sixth Annual Meet- ing, 1963, pp. 15-24. [3] — Information — A national resource, Amer. Documentation 7, 71-75 (1956). S4 0ellrich, 1963 [435], p. 72. 124 [4] Adkinson, B. W., ADI — Past, Present and Future, in P. C. Janaske [ed]. Automation and Scien- tific Communication, Pt. 3, 1964, pp. 387-388. [5] Adkinson, B. W., The Federal Government and U.S. Scientific Information, J. Chem. Documentation 2,48-50 (1962), [6] — Trends in Library Applications of Data Proc- essing, in H. Goldhor [ed.] Proceedings of the 1963 Clinic in Library Applications of Data Processing, pp. 1-8, 1964. [7] Aerospace abstracts and indexes, Special Libraries 54,117 (1963). [8] Agard Evans, A. B., International Cooperative Ab- stracting on Building : An Appraisal, in National Academy of Sciences-National Research Council, "Proceedings of the International Conference on Scientific Information, Vol. 1, 1959, pp. 491-495. [9] Alexander, S. N., 'Federal Activities in Information Interchange, in R. H. Jensen [ed]. Proceedings of Sixth Annual Meeting', 1963, pp. 67-72. [10] Alexander, S. N. and S. M. Newman, Interim Report of the Ad Hoc Interagency Study Group on Lan- guage Compatibility in Mechanized Storage and Retrieval Systems, unpublished, mimeo, June 30, 1963, 29 pp. [11] Allen, C. G., The Russian transcription of Chinese names, J. Documentation 16, 80-91 (1960). [12] Altmann, B., The Medium-Sized Information Serv- ice : Its Automation for Retrieval, Rept. No. TR- 1192. Harry Diamond Laboratories, Army Ma- teriel Command, Washington, D.C., Dec. 30, 1963, 26 pp. [13] American Association for the Advancement of Sci- ence, Information Systems — Essential Tools in Engineering Application of Science for the Needs of Society (Proceedings of a Panel Program jointly sponsored by Engineers Joint Council and Section M (Engineering) AAAS 129th Annual Meeting, Philadelphia, 1962), Engineers Joint Council, New York, 1963, 28 pp. [14] American Council of Learned Societies, Planning For Scholarly Photocopying, Preprint, Publica- tions of the Modern Language Association of America, Vol. 79, No. 4, Part 2 (Sep 64) 14 pp. [15] American Documentation Institute, Parameters of Information Science (Proceedings of Annual Meeting. Philadelphia, October 5-8, 1964), Vol. 1, Washington, D.C., 1964, 521 pp. [16] (American Federation of Information Processing Societies ) * Proceedings of the Western Joint Computer Conference, 1961, Vol. 19, Western Joint Computer Conference, Glendale, Cal., 1961, 661 pp. [17] American Federation of Information Processing Societies, Spring Joint Computer Conference, 1963, AFIPS Conference Proceedings, Vol. 23, Spartan Books, Baltimore, Md., 1963, 413 p. [18] American Institute of Chemical Engineers, Chemical Engineering Thesaurus ; A Wordbook for Use with the Concept Coordination System of Infor- mation Storage and Retrieval, New York, 1961, 175 pp. [19] American Library Association, ALA Cataloging Rules for Author and Title Entries' 2d ed., Chi- cago, 1949, 265 pp. [20] American Meteorological Society, Boston, Mass., Meteorological and Geoastrophysical Titles, Vol. 2, No. 1, Jan 1962, 55 pp. (An experimental sub- ject, author and key word index to titles listed in Part III, of Meteorological and Geophysical Abstracts, Vol. 13, No. 1, Jan 1962.) *Note that although proceedings of the Joint Computer Con- ference were not published by the American Federation of In- formation Processing Societies prior to Volume 20, they are grouped here in accordance with the volume series numbers. [21] American standard code for information inter- change, Comm. Assoc. Computing Machinery 6, 422-426 (1963). [22] American Standards Association, Inc., American Standard : Basic Criteria for Indexes, ASA Z39.4-1959, approved May 5, 1959. New York 1959, 11 pp. [23] — ■ American Standard Code for Information Inter- change, ASA X3.4-1963, approved June 7, 1943. New York, 1943, 11 pp. [24] — American Standard for Periodical Title Abbrevi- ations, ASA Z39.5-1963, approved Nov. 20, 1963. New York, 1964, 19 pp. [25] — American Standard : Reference Data and Ar- rangement of Periodicals, ASA Z39.1-1943, ap- proved June 7, 1943. New York, 1943, 4 pp. (Supersedes Z29.1-1935.) [26] — American Standard : Signaling Speeds for Data Transmission, ASA X3.1-1962, approved Aug. 8, 1962. New York, 1962, 5 pp. [27] — ■ American Standard: Specifications for 16 mm and 35 mm Microfilms on Reels or in Strips, ASA PH5.3-1958, approved Oct. 16, 1958. New York, 1959, 8 pp. [28] American University, Machine Indexing: Progress and Problems (Papers presented at the Third Institute on Information Storage and Retrieval, Feb. 13-17, 1961), Washington, D.C., 1962, 354 pp. [29] Anderson, H. T., Compatibility of Information and Data Systems Within a Company, in L. H. Hat- tery and E. M. McCormick [eds.], Information Retrieval Management, 1962, pp. 111-116. [30] Andrews, D. D., Cooperative Information — Patents, in J. H. Shera et al., [eds]. Documentation in Action, 1956, pp. 297-300. [31] Andrews, D. D. and S. M. Newman, Storage and Retrieval of Contents of Technical Literature: Nonchemical Information, Preliminary Report. Patent Office Research and Development Rept. No. 1, Washington, D.C., May 15, 1956, 12 pp. See S. M. Newman, Storage and Retrieval . . . for Supplementary Reports. [32] Angell, R. S., The Need for a New United States Code, Library Quarterly 26, 318-330 (1956). [33] — Standards for subject headings : A national pro- gram, J. Cataloguing and Classification 10, 191- 197 (1954). — Armed Services Technical Information Agency, Arlington, Va., see Defense Documentation Center. [34] Aronson, E. J., Cataloging in, a large reports or- ganization, Amer. Documentation 3, 114-117 (1952). [35] Artandi, S. and T. C. Hines, Roles and links . . . or forward to Cutter, Amer. Documentation 14, 74-77(1963). [36] Artelt, W., The problem of abbreviating the titles of medical periodicals, Bull. Med. Lib. Assoc. 43, 52-57 (1955). [37] Ashworth, W. [ed]., Handbook of Special Librarian- ship and Information Work, ASLIB, London, 1955, 387 pp. 2d ed. (completely revised), 1962, 508 pp. ASLIB Handbook 2d ed. ASLIB London, 1962, 508 pp. [38] Association of Scientific Libraries and Information Bureaus, Looking Forward in Documentation (Papers and Discussions, ASLIB 38th Annual Conference), London, 1964, 107 pp. [39] Atherton, P., A Collection of Remarks About Cita- tion Indexes, American Institute of Physics, New York, Apr 1962, 6 pp. [40] — International conference on cataloguing prin- ciples : an observer's report, Spec. Libraries 52, 583-584 (1961). [41] BA and Psychological Abstracts Commence a Program of Cooperative Abstracting, Biological Abstracts 31, x (1957). 208-371— 6G- 125 [42] Bagg, T. C. and M. E. Stevens, Information Selec- tion Systems Retrieving Replica Copies : A State- of-the-Art Report, Tech. Note 157, National Bureau of Standards, Washington, D.C., 1901, 172 pp. [43] Baker, D. B., G. M. Conrad, J. C. Green, M. Hoseh, and R. A. Jensen, Some Counterparts in Perspec- tive — a Report on Research Data Processing and Information Retrieval in Other Countries includ- ing Poland and the Soviet Union, in U.S. Con- gress, House Committee on Education and Labor, National Information Center. Hearings . . . on H.R. 1946, Appendix to Vol. 1, GPO, Wash- inton, D.C., 1963, pp. 155-372. Also available as Some Counterparts in Perspective : a Detailed Report on Visits Oct. 20 to Nov. 5, 1959, National Federation of Science Abstracting and Indexing Services, Washington, D.C., 1960, lv. [44] Ball, N. T., Research or available knowledge : a matter of classification, Science 105, 34-36 (1947). [45] Bar-Hillel, Y., Is information retrieval approaching a crisis? Amer. Documentation 14, 95-98 (1963). [46] Becker, J., Machines and Equipment, in H. Orlans [ed], Federal Departmental Libraries: a Sum- mary Report of a Survey and a Conference, 1963, pp. 87-93. [47] — Some Approaches to Mechanization of Technical Information Processing Systems, in D. E. Cetone [ed]. Proceedings of the March AFBMD Confer- ence on Scientific and Technical Information, 1960, pp. 9-20. [48] Becker, J. and R. M. Hayes, Information Storage and Retrieval : Tools, Elements, Theories, 448 pp. (John Wiley & Sons, New York, N.Y., 1963), 44S p. [49] Bemer, R. W., The American standard code for in- formation interchange, parts 1-2, Datamation 9, 32-36 (Aug. 1963) and 9, 39-41, 44 (Sep. 1963). [50] — The Present Status, Achievement and Trends of Programming for Commercial Data Processing, in W. Hoffmann [ed]. Digital Information Proces- sors, 1960, pp. 312-405. [51] — Survey of coded character representation, Comm. Assoc. Computing Machinery 3, 639-642 (1960). [52] Bennington, J., The integration of report literature and journals, Amer. Documentation 3, 149-152 (1952). [53] Bernal, J. D., Provisional Scheme for Central Dis- tribution of Scientific Publications, in Royal Society of London. Royal Society Scientific Infor- mation Conference, London, June 21-July 2, 1948, pp. 253-258. [54] — The Transmission of Scientific Information : a User's Analysis, in National Academy of Sciences- National Research Council, International Con- ference on Scientific Information, Vol. 1, 1959, pp. 77-95. [55] Bernier, C. L. and E. J. Crane, Correlative indexes, part VIII. Subject-indexing versus word-index- ing, J. Chem. Documentation 2, 117-122 (1962). [56] Bernier, C. L. and K. L. Heumann, Correlative indexes, part III. Semantic relations among semantemes — the technical thesaurus, Amer. Documentation 8, 221-220 (1957). [57] Bernstein, H. H, The use of flexowriters in docu- mentation centers and libraries, UNESCO Bull. Libraries 16, 79-85 (1962). [58] Bernstein, H. H. and K. H. Meyer-Uhlenried, Re- cording Bibliographical Data by Punching Type- writers — A Survey, in H. P. Luhn fed]. Automation and Scientific Communication, Pt. 1, 1963, pp. 75-76. [59] Berry, M. M. and J. W. Perry, Notational systems for structural formulas, Chem. Eng. News 30, 407 (1952). 126 [60] Bessinger, J. B. Jr., S. M. Parrish, and H. F. Arader [eds], Literary Data Processing Conference Proceedings September 9, 10, 11, 1964, IBM, White Plains, N.Y., 1964, 329 pp. [61] Better Information Coordination Urged, Chem. Eng. News 42, 27 (1964). [62] Bishop, C, An integrated approach to the docu- mentation problem, Amer. Documentation 4, 54- 65 (1953). [63] Biunno, V. P., Cooperative Information Process- ing — Patents, in J. H. Shera et al. [eds], Docu- mentation in Action, 1956, pp. 359-364. [64] Blegen, T. O, The scholar looks at inter-library co- operation, Lib. Quarterly 22, 13-17 (1952). [65] Bliss, H. E., The Organization of Knowledge and the System of the Sciences, 433 pp. (Holt, New York, 1929). [66] Boaz, M. [ed], Modern Trends in Documentation, 103 pp. (Pergamon Press, New York, 1959). [67] Bohr, P. G. and K. C. Owen. A cooperative project in new drug reporting, J. Chem. Documentation 3,201-205 (1963). [68] Bonn, G. J., Training for Activity in Scientific Docu- mentation Work, 1959, pp. 1441-1488, in National Academy of Sciences — National Research Coun- cil, Proceedings of the International Conference on Scientific Information, Vol. 2, 1959, pp. 1441- 1488. [69] Born, L. K., The Literature of Microreproduction, 1950-1955, Amer. Documentation 7, 167-187 (1956). [70] — Universal Guide to catalogs of manuscripts and inventories of archival collections : a proposal for cooperative listing, College and Research Li- braries 17, 322-329 (1956). [71] Bourne, C. P., The historical development and pre- sent state-of-the-art of mechanized information retrieval systems, 241 pp. Amer. Documentation 12,108-110 (1961). [72] — Methods of Information Handling (John Wiley & Sons, New York, N.Y., 1963). [73] — The world's technical journal literature: an estimate of volume, origin, language, field, in- dexing, and abstracting, Amer. Documentation 13,159-168 (1962). [74] Boutry, G. A., Abstracting Board of International Council of Scientific Unions, Science 123, 423-424 (1956). [75] Boyd, J. P., A landmark in the history of coopera- tion, College and Research Libraries 8,' 101-109 (1947). [76] Bralove, A. L., Concept, Specifications, Dissemina- tion, and User Acceptance Study of the NASA Microform, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 1, 1963, pp. 77-78. [77] Brandhorst, W. T., Corporate author cataloging and the technical report literature, Amer. Documenta- tion 15,35-46 (1964). [78] Brenner, C. W. and C. N. Mooers, Case History of a Zatocoding Retrieval System in R. S. Casey et al. [eds], Punched Cards: Their Application to Sci- ence and Industry, 1958, pp. 340-356. [79] Brenner, E. H, B. H. Weil and N. E. Rawson, A New Centralized Information-Retrieval System for the Petroleum Industry Including a Computer Search System and Two Manual Indexes, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 2, 1963, pp. 291-292. [80] A Brief Description of Viniti (with illustrations) — the Soviet Union's All Union Institute of Scien- tific and Technical Information, located in Moscow, May 1963, in U.S. Congress, House Com- mittee on Education and Labor, National Infor- mation Center. Hearings ... on H.R. 1946, Appendix to Vol. 1, 1963, pp. 373-409. [81] British Standards Institution, Guide to the Uni- versal Decimal Classification (UDC), B.S. lOOOC, London, 1963, 128 pp. (F.I.D. No. 345). [82] — Universal Decimal Classification. Abridged English edition, B.S. 1000A, London, 1961, 254 pp. (F.I.D. No. 289). [83] Browne, C. A. and M. B. Weeks. A History of the American Chemical Society, American Chemical Society, Washington, D.C., 1952, 526 pp. [84] Brownson, H. L., Current United States Research on Topics Related to the Conference, in Interna- tional Federation for Documentation, Proceed- ings of the International Study Conference on Classification for Information Retrieval, 1957, pp. 99-100. [85] — Documentation Needs of Scientists, in L. H. Hattery and E. M. McCormick [eds], Information Retrieval Management, 1962, pp. 76-82. [86] — Recommendations and results of international conferences on scientific information and biblio- graphic services, Amer. Documentation 3, 29-55 (1952). [87] Buckland, L. F., Recording Text Information in Machine Form at the Time of Primary Publica- tion, in H. P. Luhn [ed], Automation and Scien- tific Communication, Pt. 2, 1963, pp, 309-310. [88] Burnett, M. M., The Interservice Data Exchange Program as a Specialized Approach, in D. E. Cetone [ed], Proceedings of the March AFBMD Conference on Scientific and Technical Informa- tion, 1960, pp. 31-34. [89] Burns, L. R., Automation in The Public Libraries of Lake County, Indiana, in H. Goldhur [ed], Proceedings of the 1963 Clinic on Library Appli- cations of Data Processing, 1964, pp. 9-17. [90] Bush, V., As we may think, Atlantic Monthly 176, 101-108 (1945). [91] — Science the Endless Frontier, A Report to the President on a Program for Postwar Scientific Research, U.S. Government Printing Office, Wash- ington, D.C., July 1945, 184 pp. [92] Cady, W. H., An examination of scientific periodi- cals, Standardization 21, 11-12 (1950). [93] Cahn, J. N., A B.I.T.S.— Bureau of Information for Technology and Science proposed as Key Station of Federal Network of Information Services to U.S. Free Enterprise Systems, in Engineers Joint Council, Proceedings of the EJC Engineering Information Symposium, 1962, pp. 45-50. [94] ■ — ■ A System of Information Systems, in L. H. Hattery and E. M. McCormick [eds], Informa- tion Retrieval Management, 1962, pp. 21-30. [95] Cambridge Language Research Unit, England. Re- port on Research, Mech. Translation 3, 36-37 (1956). [96] Canter, L., Factors restricting the availabilty of Government research reports, Amer. Documenta- tion 3, 133-138 (1952). [97] Carlson, I. G., Cataloging of technical reports in a Military Research Library, Amer. Documenta- tion 3, 119-122 (1952). [98] Carlson, W. M., Information Resources for Tomor- row's Engineers, in American Association for the Advancement of Science, Information Systems — Essential Tools in Engineering Application of Science for the Needs of Society, 1963, pp. 21-28. [99] — Proposed EJC Course of Action in Field of Information Systems, Engineers Joint Council, New York, 21 Sep. 1962, 13 pp., appendices. [100] Casey, R. S., J. W. Perry, M. M. Berry, and A. Kent [eds], Punched Cards: Their Applications to Science and Industry, 697 pp. (Reinhold Pub- lishing Corp., New York, N.Y., 2nd ed., 1958). [101] Cataloging-in-Source — a Symposium, Library Res. Tech. Services 4, 269-284 (1960). [102] Center for Documentation and Communication Re- search, Western Reserve University, Information Retrieval in Action (Proceedings of Conference on Research in Progress 1962), The Press of Western Reserve University, Cleveland, Ohio, 1963, 323 p. [103] Center for text in machine-usable form, Scientific Information Notes 5, 1 (1963) . [104] Centralized register of microfilms, Library of Congress Inf. Bull. 22, 593-594 (1963). [105] Cetone, D. E. [ed], Proceedings of the March AFBMD Conference on Scientific and Technical Information, Rept. No. AFBMD TR-60-143, Air Force Ballistic Missile Division, Air Research and Development Command, Washington, D.C., 1960, 134 pp. [106] Chamberlin, W., An International Institute for Scientific Information, in National Academy of Sciences — National Research Council, Proceed- ings of the International Conference on Scientific Information, Vol. 2, 1959, pp. 1523-1534. [107] Chaplin, A. H., Co-operative cataloging, Library Assoc. Record 55, 275-278 (1953). [108] — International Conference on cataloguing princi- ples. I. Organization, J. Documentation 19, 41- 46 (1963). [109] — A universal cataloging code, Lib. Quarterly 26, 337-347 (1956). [110] Chemical Center. Clearing House will screen all organic compounds to aid in research for new drugs, insecticides, Federal Science Progress 1, 15 (1947). [Ill] Cheydleur, B. F., Information retrieval— 1966, Data- mation 7,21-25 (1961). [112] Chollar, R. G., Standards in computers and infor- mation processing, J. Data Management 1, 38-42 (1963). [113] Clapp, V. W., Cooperative Acquisitions, College and Research Libraries 8, 99-100 (1947). [114] — Libraries and the 'Uppercase Limitation', in B. E. Markuson [ed], Libraries and Automation, 1964, pp. 54-55. [115] — The purchase of books in Europe, College and Research Libraries 7, 127-134 (1946). [116] — Research in Problems of Scientific Information — Retrospect and prospect, Amer. Documentation 14,1-10 (1963). [117] — ■ The Role of Bibliographic Organization in Con- temporary Civilization, in J. H. Shera and M. E. Egan [eds], Bibliographic Organization, 1951, pp. 3-23. [118] — Standards in Photographic Reproduction of Documents, Library of Congress Inf. Bull. 12, 12- 13 (1953). [119] — Subject controls — Nature and levels of controls, Amer. Documentation 3, 11-15 (1952). [120] — UNESCO's Program : Libraries, Bibliographic Services, Documentation, A Catalog of Principal Activities, Amer. Documentation 7, 127-134 (1956). [121] Clapp, V. W. and F. B. Rogers, Responsibilities of Government Professional Societies, Universities, and Industry for Improved Information Services and Research, Summary of discussion for Area 7, in National Academy of Sciences-National Re- search Council, Proceedings of the International Conference on Scientific Information, Vol. 2, 1959, pp. 1535-1545. [122] Cleverdon, C. W., The ASLIB Cranfield Research Project on the Comparative Efficiency of Index- ing Systems, ASLIB Proc. 12, 412-431 (1960). [123] — The Evaluation of Systems Used in Information Retrieval, in National Academy of Sciences-Na- tional Research Council, Proceedings of the In- ternational Conference on Scientific Information, Vol. 1, 1959, pp. 687-698. [124] — Report on the Testing and Analysis of an In- vestigation into the Comparative Efficiency of In- dexing Systems, College of Aeronautics, Cranfield. England, Oct. 1962, 311 pp. [125] Cleverdon, C. W., F. W. Lancaster and J. Mills, Un- covering some facts of life in information re- trieval, Spec. Libraries 55, 84-91 (1964). 127 [126] Cleverdon, C. W. and J. Mills, The Analysis of In- dex Language Devices, in P. C. Janaske [ed], Automation and Scientific Communication, Pt. 3, 1964, pp. 451-454. [150] [127] — The Testing of Index Language Devices, ASLIB Proc. 15, 106-130 (1963), [128] Coblans, H., International and Local Information Services, in Library Association, Gt. Britain, Scarborough Conference, Proceedings . . . Dis- [151] cussions, 1960, pp. 67-73. [129] Collison, R. L., Bibliographical Services Through- out the World 1950-59, UNESCO, Paris, 1961, 228 pp. [152] [130] — The Continuing Barrier. Translations as a Factor of East- West Communications : an Analy- sis of the Index Translationum, UNESCO Bull. [153] Libraries 16, 296-300 (1962). [131] — Indexes and Indexing: Guide to the Indexing of Books, and Collections of Books, Periodicals, [154] Music, Gramaphone Records, Films, and Other Material, with a Reference Section and Sugges- tions for Further Reading, 200 p. (de Graff, New [155] York, 1959). [132] Compton, A. H., UNESCO and science, Federal Sci- ence Progress 1, 21-23 (1947). [133] Conferences on Science and World Affairs. State- [156] ments by Participants at the 7th and 8th Confer- [157] ences, at Stowe, Vt., in September. Science 184, 984-991 (1961). [134] Connor, J. M., The need for documentation to Gov- ernment specifications, Spec. Libraries 47, 152- 155 (1956). [158] [135] Cook, C. D., The Farmington plan and the select list of unlocated research books, College and Re- search Libraries 15, 281-284, 312 (1954). [136] Cornelius, M. E., Machine Input Problems and Ma- chine Indexing Alternatives and Practicalities, in [159] American University, "Machine Indexing: Prog- ress and Problems", 1962, pp. 41-49. [160] [137] Costello, J. C. Jr., Computer requirements for in- verted coordinate indexes, Amer. Documentation 13,414-419 (1962). [138] — Uniterm indexing principles, problems and solu- tions, Amer. Documentation 12, 20-26 (1961). [161] [139] Council on Library Resources, Inc., First Annual Report for the period ending June 30, 1957, Wash- ington, D.C., 1957, 28 pp. [162] [140] — Second Annual Report for the period ending June 30, 1958, Washington, D.C., 1958, 43 pp. [163] [141] — Third Annual Report for the period ending June 30, 1959, Washington, D.C., 1959, 63 pp. [164] [142] — Fourth Annual Report for the period ending June 30, 1960, Washington, D.C., 1960, 44 pp. [165] [143] — Fifth Annual Report for the period ending June 30, 1961, Washington, D.C., 1961, 66 pp. [166] [143] — Fifth Annual Report for the period ending June 30, 1961, Washington, D.C., 1961, 66 pp. [167] [144] Council on Library Resources, Inc., Library of Con- gress Sponsors Experimental Demonstration Involving Use of Tape-Producing Typewriter for [168] Versatile Automated Printing, Recent Develop- ments No. 129, Washington, D.C., May 4, 1964, 3 [169] pp. [145] Cox, G. J., C. F. Bailey and R. S. Casey. Punch Cards [170] for a Chemical Bibliography, Chem. Eng. News 23, 1623-1626 (1945). [146] Craig, G. Y., Abstract Cards in Scientific Journals, [171] in Association of Scientific Libraries and Informa- tion Bureau, Looking Forward in Documentation, 1964, pp. 4-12 to 4-14. [172] [147] Crane, E. J. and C. L. Bernier, Indexing and Index Searching, in R. S. Casey et al. [ed], Punched Cards . . . 1958, pp. 510-527. [148] Creating order out of confusion, Business Automa- [173] tion 11, 22-26 (July 1964) . [149] Crestadoro, A., The Art of Making Catalogues of [174] Libraries : A Method to Obtain in a Short Time a Most Perfect, Complete, and Satisfactory Cata- logue of the British Museum Library, By a Reader Therein, The Literary, Scientific and Artistic Reference Office, London, 1856, 60 pp. Crowther, B. M., International Cooperation in Physics Abstracting, in National Academy of Sciences — National Research Council, Proceed- ings of the International Conference on Scientific Information, Vol. 1, 1959, pp. 481-489. Croxton, F. E., Identification of Technical Reports, in B. M. Fry and J. J. Kortendick [eds], The Production and Use of Technical Reports, 1955, pp. 121-135. Cunningham, R., International Congress of Librar- ies and Documentation Centers, Spec. Libraries 47,32-34 (1956). Cutter, C. A., Rules for a Dictionary Catalog, 4th ed., U.S. Government Printing Office, Washing- ton, D.C., 1904. Danton, E. M., [ed], The Library of Tomorrow: a Symposium, American Library Association, Chi- cago, 1939, 191 pp. Datatrol Corp., Common Vocabularly Approaches for Government Scientific and Technical Informa- tion Systems, Technical Rept. No. IR-10, Contract NSF C-342, Silver Spring, Md., 1963, 108 pp. Davies, D. W., Rev. Int. Doc. 29, 66-68 (1962). Davis, D. D., Subject Indexing for Nuclear Science Abstracts, in Atomic Energy Commission, in U.S. Atomic Energy Commission, The Literature of Nuclear Science: Its Management and Use, 1962, pp. 209-214. — The Use of Punched-Tape Typewriters and Com- puters in the Centralized Information Processing at the USAEC Division of Technical Information Extension, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 2, 1963, pp. 237-238. Davis, W., Need information center, Science News Letter 84, 70-71 (1963). Day, M. S., The Scientific and Technical Informa- tion Program of the National Aeronautics and Space Administration, in U.S. Atomic Energy Commission, The Literature of Nuclear Science : Its Management and Use, 1962, pp. 361-377. Day, M. S. and I. Lebow, New indexing pattern for nuclear science abstracts, Amer. Documentation 11,120-127 (1960). Defense Documentation Center, ASTIA Chemical Thesaurus, Alexandria, Va., Dec. 1962, 34 pp. — ■ ASTIA Guidelines for Cataloging & Abstracting, Alexandria, Va., June 1962. 29 pp. — Radiology, Information Service Series, Vol. 2, No. 1, Alexandria, Va., Apr. 1963, 255 pp. — Technical Abstracts Bulletin Cumulative Index, Jan.-June 1962, Alexandria, Va., 1962, 1007 pp. — Technical Abstract Bulletin Cumulative Index, July-Dec. 1962, Alexandria, Va., 1963, 805 pp. — Thesaurus of ASTIA Descriptors, 1st ed., Alexandria, Va., May 1960, 328 pp., 2d ed., Dec. 1962, 673 pp. Defense research reports indexed, Scientific Infor- mation Notes 5, 13 (Dec. 1963- Jan. 1964). De Grolier, E., Problems in scientific communica- tion, IBM J. Res. Dev. 2, 276-281 (1958). — A Study of General Categories Applicable to Classification and Coding in Documentation, UNESCO, Paris 1962. 248 pp. Deignan, S. L. and E. Miller, The support of research in medical and allied fields for the period 1946 through 1951, Science 115, 321-343 (1952). Dennis, B. K., Financing a Technical Information Center, in L. H. Hattery and E M. McCormick [eds], Information Retrieval Management, 1962, pp. 61-75. Dewey, H., Punched card catalogs — theory and tech- nique, Amer. Documentation 10, 36-50 (1959). Dewey, M., Dewey Decimal Classification and Rela- tive Index, 16th ed., Lake Placid Club, New York, Forest Press, 1958. 2 v., 2439 pp. 128 [175] Distribution, Federal Science Progress 1, 10 (1947). [176] Donker Duyvis, F., Standardization as a tool of scientific management, Library Trends 2, 410-427 (1954). [177] Dougherty, R. M., The Scope and Operating Effi- ciency of Information Centers as Illustrated by the Chemical-Biological Coordination Center of the National Research Council, Ph. D. Disserta- tion, Rutgers-The State University, New Bruns- wick, N.J., May 1963, 241 pp. [178] Downs, R. B., Latin American union catalogs, Col- lege and Research Libraries 7, 210-213 (1946). [179] — One for All: a Historical Sketch of Library Co-operation, 1930-1970, in E. M. Danton [ed], The Library of Tomorrow : a Symposium, 1939, pp. 60-67. [ISO] — Wartime cooperative acquisitions, Lib. Quarterly 19,157-165 (1949). [181] Dubester, H., The Role of the American Documen- tation Institute in International Documentation, Amer. Documentation 13, 115-117 (1962). [182] Dyson, G. M. and J. E. L. Faradane, The aims of the Institute of Information Scientists Ltd., J. Chem. Documentation 2, 72-74 (1962). [183] Edwards, P. I., Information Retrieval in the Bi- ological Sciences, in Association of Scientific Libraries and Information Bureaus, Looking Forward in Documentation, 1964, pp. 4-15 to 4-18. [184] Electronic and other equipment for chemical infor- mation work, Chem. Eng. News 27, 2911 (1949). [185] Engineers Joint Council, Proceedings of the EJC Engineering Information Symposium, New York, 1962, 95 pp. [186] — Study of Engineering Terminology and Relation- ships Among Engineering Terms, Final report on NSF Grant G 24836, New York, Aug 1963, 16 pp. [187] Evans, L. H., History and the problem of bibliogra- phy, College and Research Libraries 7, 195-205 (1946). [188] — Survey of Federal Departmental Libraries, in H. Orlans [ed], Federal Departmental Libraries: A Summary Report of a Survey and a Conference, 1963, pp. 3-54. [189] — UNESCO work and method illustrated by the library programs, Lib. Quarterly 24, 92-100 (1954). [190] Fagerhough, K. H., Availability of industrial re- search reports, Amer. Documentation 3, 143-145 (1952). [191] Fairthorne, R. A., Implications of Test Procedures, in Center for Documentation and Communication Research, Information Retrieval in Action, 1963, pp. 109-113. [192] Fall, J., Problems of American libraries in acquir- ing foreign publications, Lib. Quarterly 24, 101- 113 (1954). [193] Farradane, J., Introduction, Inf. Storage and Retrieval 1,1 (1963). [194] — Relational indexing and classification in the light of recent experimental work in psychology, Inf. Storage and Retrieval 1, 3-11 (1963). [195] Federal Council for Science and Technology. Com- mittee on Scientific Information, Standard for Descriptive Cataloging of Government Scientific and Technical Reports, Washington, D.C., Dec 1963, 21 pp. [196] — Status Report on Scientific and Technical In- formation in the Federal Government, Washing- ton, D.C., June 18, 1963, 18 pp. [197] Field, H. W., A Centralized Abstracting Service for the Petroleum Industry, in J. H. Shera et al. [eds], Documentation in Action, 1956, pp. 269-278. [198] Focke, H. M., Cooperative Information Processing — Documentation, in J. H. Shera et al. [eds], Docu- mentation in Action, 1956, pp. 355-358. [199] Foskett, D. J., The Classification Research Group, 1952-1962, Libri 12, 127-138 (1962). [200] Frank, O., Cooperation and Coordination in Ab- stracting and Documentation, in National Acad- emy of Sciences-National Research Council, Proceedings of the International Conference on Scientific Information, 1959, Vol. 1, pp. 497-510. [201] Frarey, C. J., The role of research in establishing standards for subject headings, J. Cataloging and Classification 10, 179-190 (1954). [202] — Subject Headings, in R. R. Shaw [ed], The State of the Library Art, Vol. 1, Pt. 2, 1960, pp. 5-67. [203] Freeman, M. E., Technical information for research program management, J. Chem. Documentation 3,217-220 (1963). [204] Freeman, R. R., Computers and Classification Sys- tems, J. Documentation 20, 137-143 (1964). [205] Frontard, R., Transliteration codes and their international standardization, UNESCO Bull. Libraries 15, 78-82 (1961). [206] Fry, B. M., and J. J. Kortendick [eds], The Produc- tion and Use of Technical Reports, (Proceedings of the Workshop on the Production and use of Technical Reports, Conducted at the Catholic University of America, 13-18 Apr 1953), The Catholic University of America Press, Washing- ton, D.C., 1955, 175 pp. [207] Fussier, H. H., The Problems of Physical Accessi- bility, in J. H. Shera and M. E. Egan [eds], Bibliographic Organization, 1951, pp. 163-186. [208] — Readjustments by the librarian, Lib. Quarterly 23,216-229 (1953). [209] Gardin, J. C, Document analysis and information retrieval, UNESCO Bull. Libraries 14, 2-5 (1960). [210] Garfield, E., Citation indexes for science: A new dimension in documentation through association of ideas, Science 122, 108-111 (1955). [211] — A Unified Index to Science, in National Academy of Sciences — National Research Council, Proceed- ings of the International Conference on Scientific Information, Vol. 1, 1959, pp. 461-474. [212] Gelberg, A., W. Nelson, G. S. Yee, and E. A. Metcalf, A program retrieval of organic struc- ture information via punched cards, J. Chem. Documentation 2, 7-11 (1962). [213] General principles of cooperation on biological ab- stracting, Science 123, 578 (1956). Also in Bio- logical Abstracts 30, viii (1956). [214] Genereaux, R. P., Technical information needs of the chemical processing industry, J. Chem. Documentation 2, 40-45 (1962). [215] Gibb, I. P., Foreign book procurement: the decen- nial Farmington plan survey and afterwards, J. Documentation 16, 1-9 (1960). [216] Gillett, H. W., Metallurgical Research as a Na- tional Resource, in National Resources Planning Board, Research — A National Resource II. In- dustrial Research, 1941, pp. 289-305. [217] Gingold, K., Translation Pools-ideal and reality, J. Chem. Documentation 1, 14-19 (1961). [218] Goldhor, H. [ed], Proceedings of the 1963 Clinic on Library Applications of Data Processing, Grad- uate School of Library Science, University of Illinois, Champaign, 111., 1964, 176 pp. [219] Government microfiche programs now shifting into high gear, Systems 4, 26-28 (1963). [220] Gray, B. S., Automatic typing of literature searches, J. Chem. Documentation 1, 71-73 (1961). [221] Gray, D. E., An experiment in standardization, Physics Today 4, 8-9 (1951). [222] — Office of Technical Services of the Department of Commerce, Physics Today 4, 24-26 (1951). [223] — Practical experience in microfacsimile publica- tion, Amer. Documentation 3, 58-61 (1952). [224] Green, J. C, Last call for Germany, Federal Science Progress 1, 24-25 (1947). [225] — The Role of the Department of Commerce, J. Chem. Documentation 3, 223-226 (1963). 129 [226] — Statement, in U.S. Congress, House Committee [249] on Science and Astronautics, Dissemination of Scientific Information, 11)51), pp. 11(5-122. [227] Greins, M., Survey of Languages and Systems for Information Retrieval, IBM Research Center, Yorktown Heights, New York, 1961. [228] Gull, C. 1)., Automatic documentation, current sys- tems and trends in the USA, Rev. Int. Doc. 29, 57-62 (1962). [229] — Guidelines to Mechanizing Information Systems, in L. H. Hattery and E. M. McCormick [eds], In- formation Retrieval Management, 1962, pp. [250] 101-110. [230] — The Impact of Electronics Upon Cataloguing Rules, in International Federation of Library Associations, "International Conference on Cata- loguing Principles: Report", 1963, p. 281-290. [251] [231] — Seven Years of Work on the Organization of Ma- terials in the Special Library, Amer. Documenta- [252] tion 7, 320-329 (1956). [232] Giinther, A., Microphotography in the Library, UNESCO Bull. Libraries 16, 1-22 (1962). [233] Hall, J. H., The Idea : The Dissemination Manage- ment Concept, in D. E. Cetone [ed], "Proceed- [253] ings of the March AFBMD Conference on Scien- tific and Technical Information", 1960, p. 61-63. [234] Halliday, M. A. K., The Linguistic Basis of a Me- [254] chanical Thesaurus, Mech. Translation 3, 81-88 (1956). (Paper presented at the Second Inter- national Conference on Machine Translation, held at the Massachusetts Institute of Technology, [255] Cambridge, Mass., Oct 16-20, 1956). [235] Hammond, W., Convertibility of Indexing Vocab- ularies, in U.S. Atomic Energy Commission, The [256] Literature of Nuclear Science : Its Management and Use. 1962, p. 223-234. [236] Hammond, W. and S. Rosenborg, Experimental Study of Convertibility Between Large Technical [257] Indexing Vocabularies, Technical rept. no. IR-1, Contract no. NSF C-259. Datatrol Corp., Silver Spring, Md., Aug 1962, 297 p. [258] [237] Hamrick. L. A., The Role of the Office of Technical Services in the Interchange of Documents, in H. P. Luhn [ed]. Automation and Scientific Com- munication, Pt. 2, 1963, p. 219-220. [259] [238] Hanson, C. W. and M. Janes, Coverage by Abstract- ing Journals of Conference Papers, J. Documenta- tion 17, 143-149 (1961). [260] [239] Haskins, S. M., Moving Toward International Cata- loging Agreement, ALA Bull. 54, 197-201 (1960). [240] Hattery, L. H. and E. M. McCormick [eds], Infor- [261] mation Retrieval Management, (Papers adapted from Proceedings of the American University [262] Fourth Institute on Information Storage and Re- trieval 1962), American Data Processing, Inc., Detroit, Mich., 1962, 151 p. [241] Hayes, R. M., The Integrated Systems Approach, in [263] D. E. Cetone [ed], Proceedings of the March AFBMD Conference on Scientific and Technical Information, 1960, p. 5-7. t 264 l [242] Heald, J. H, ASTIA and the Information Revolu- tion, Spec. Libraries 54, 40-44 (1963). [243] — Project MARS, Spec. Libraries 51, 115-121 [265] (1960). [244] — The Reports Librarian and National Security, Amer. Documentation 8, 138-143 (1952). [245] - — Some Search Strategy Factors in Automated Searches of Technical Reports at DDC, in H. P. [266] Luhn [ed], Automation and Scientific Communi- cation, Pt. 2, 1963, p. 159-160. [267] [246] Heilprin, L. B., Communication Engineering Approach to Microforms, Amer. Documentation [268] 12,213-218 (1961). [247] — On the Information Problem Ahead, Amer. Docu- mentation 12, 6-14 (1961). [248] Heising, W. P., FORTRAN: Compatibility and Standardization, Datamation 10, 24-25 (Aug [269] 1964). 130 Heller, Robert & Associates, Inc., Cleveland, Ohio, A National Plan for Science Abstracting and Indexing Services, National Federation of Science Abstracting and Indexing Services, Washington, D.C., NSF Grant G22594, Mar 15, 1963, 38 p. Also in R. A. Jensen [ed], "Proceedings of Sixth An- nual Meeting", 1963, Appendix A, 38 p. ; and U.S. Congress, House Committee on Education and Labor, "National Information Center, Hearings . . . on H. R. 1946", Appendix to Vol. 1, 1963, p. 100-153. Henkle, H. H., The Natural Sciences : Character- istics of the Literature, Problems of Use, and Biblographic Organization in the Field, in J. H. Shera and M. E. Egan [eds], Bibliographic Orga- nization, 1951, p. 140-160. Henriques, V., The Urgent Quest for Standards, Business Automation 11, 22-27 (1964). Hensley, C. B., Selective Dissemination of Informa- tion (SDI) : State of the Art in May 1963, in American Federation of Information Processing Societies, Spring Joint Computer Conference 1963, p. 257-262. Herbert, E., "Finding What's Known", Interna- tional Science and Technology, No. 1, Jan. 1962, p. 14-23. Herner, M. and S. Herner, "The Current Status of the Government Research Report in the United States of America", UNESCO Bull. Lib. 13, 187- 196 (1959). Herner, S., Methods of Organizing Information for Storage and Searching, Amer. Documentation 13, 3-14 (1962). — "The Role of Theasuri in the Convergence of Word and Concept Indexing", in H. P. Luhn [ed], "Automation and Scientific Communication, Pt. 2, 1963, p. 183-184. — The Storage and Housing of Unpublished Research Reports, Amer. Documentation" 3, 110- 114 (1952). Heumann, K. F., International Activities in Docu- mentation, in L. H. Hattery and E. M. McCormick [eds], "Information Retrieval Management", 1962, p. 120-122. — The Program of the United States National Committee for FID, Amer. Documentation 14, 78- 81 (1963). Hicks, M. S., Government-Sponsored Research Re- ports in Three Areas of Physical Chemistry, J. Chem. Documentation 3, 144-148 (1963). Hillier, J., Measuring the Value of Information Services, J. Chem. Documentation 2, 31-34 (1962). Hilsenrath, J., The Utility of Technical Reports, in B. M. Fry and J. J. Kortendick [eds], The Produc- tion and Use of Technical Reports, 1955, p. 104-110. Hilty, D. C, Cooperative Information Processing — Metals, in J. H. Shera et al. [eds], "Documenta- tion in Action", 1956, p. 341-342. Hines, T. C. The Collectanea as a Bibliographic Tool, PhD. Thesis (Rutgers) 1961, University Microfilms, Ann Arbor, Mich., 1961. 233 p. — Machine Arrangement of Alphanumeric Concord- ance, Thesaurus, and Index Entries : the Need for Compatible Standard Rules, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 1, 1963, p. 7-8. Hoffmann, W. [ed], Digital Information Processors, Interscience, New York, 1962, 740 p. Holloway, A. H., The Price of Co-operation, ASLIB Proe. 13, 15-25 (1961). Holm, B. E., Information Retrieval — a Solution, Chem. Eng. Progress 57, 73-78 (1961). Also in Engineers Joint Council, "Proceedings of the EJC Engineering Information Symposium", 1962, p. 85-90. Holmes, D. C, Photoduplication of Technical Reports, in B. M. Fry and J. J. Kortendick [eds], The Production and Use of Technical Reports, 1955, p. 50-59. [270] Holstrom, J. B., How Translators Can Contribute to Improving Scientific Terminology, Babel 1, 73- 79 (1955). [271] Hooker, R. M., Manual System at NRL Works Well, Survey Advises Against Mechanization, Navy Management Review 6, 10-11, 17 (1962). [272] Horecky, P. L., Government Publications, American Association for the Advancement of Slavic Studies, Newsletter, Fall 1960, p. 17-18. [273] Horty, J. F. and T. B. Walsh, Use of Flexowriters to Prepare Large Amounts of Alphabetic Legal Data for Computer Retrieval, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 2, p. 259-260. [274] Howerton, P. W., Status of Technical Information Centers, in L. H. Hattery and E. M. McCormick [eds], Information Retrieval Management, 1962, p. 31-33. [275] Humphrey, H. H., Defense Information on Science and Technology ; Key to Progress in National Research, Congressional Record 109, 18124-18129 (1963). [276] Humphrey, R. A., The Department of State and the Acquisition of Research Materials of Foreign Origin, College and Research Libraries 7, 99-104, 108 (1946). [277] Hutchisson, E., Cooperative Information Processing — Science Abstracting, in J. H. Shera et al. [eds], Documentation in Action, 1956, p. 279-282. [278] Hyslop, M. R.. A Compatibility Study of Two Information Systems, Amer. Documentation 14. 292-298 (1963). [279] — Cooperative Information Processing in Metals, in J. H. Shera et al. [eds], "Documentation in Action", 1956. p. 257-266. [280] — International Cooperation in Metallurgical Lit- erature Classification, Amer. Documentation 4, 1-3 (1953). [281] Intermittent Publications, Science 125, 487-488 (1957). [282] International Conference on Science Abstracting, 1949, Final Report, UNESCO, Paris, 1949, 1 v. [283] International Congress on Medical Librarianship, 1st, London, July 20-25. 1953, Proceedings, Libri 3, 9-16, 258-278, 285-294, 402-418 (1954). [284] International Federation for Documentation, Mod- ern Documentation and Information Practices, The Hague, 1961, 225 p. [285] International Federation for Documentation. Pro- ceedings of the International Study Conference on Classification for Information Retrieval, ASLIB, London ; Pergamon, New York, 1957, 151 p. [286] International Federation of Library Associations, International Conference on Cataloguing Prin- ciples: Preliminary Official Report, UNESCO Bull. Libraries 16, 53-63 (1962). [287] — International Conference on Cataloguing Prin- ciples: Report, (held in Paris, Oct 9-18, 1961), London, 1963, 293 p. [288] International Organization for Standardization, Standardizaton in the Domain of Documentation, The Hague, 1954, 60 p. [289] International Union of Pure and Applied Chem- istry, A Proposed International Chemical Nota- tion, Tentative Version Subject to Revision, Longmans, Green, New York, 1958, 165 p. [290] Italy-United States Exchange Manuscripts on Paper Tapes, Scientific Information Notes 5, 1 (1963). [291] Jackson, E. B., Acquisitions: Sources and Tech- niques, Amer. Documentation 3, 94-100 (1952). [292] — The NACA Classification Svstem— a Case His- tory, ASLIB Proc. 5, 215-234 (1953). [293] Jacoby, J. and V. Slamecka, Indexer Consistency Under Minimal Conditions, Contract AF 30 (602) 2616, Documentation, Inc., Bethesda, Md., Nov 1962, 1 v. [294] Janaske, P. C. [ed], Automation and Scientific Com- munication, (Proceedings of the 26th Annual Meeting of the American Documentetis Institute, Oct 6-11, 1963), American Documentation Insti- tute, Washington, D.C., 1964, Pt. 3, p. 385-489. [295] Jaster, J. J., A Note on Descriptors, Amer. Docu- mentation 13, 433-434 (1962). [296] — Subsumption Scheme for Dictionary of Indexing Equivalents, Technical rept. No. IR-3, Contract NSF C-259, Datatrol Corp., Silver Spring, Md., Feb 1963, 1 v. [297] Jaster, J. J., B. R. Murray and M. Taube, The State of the Art of Coordinate Indexing, Contract NSF C-147, Documentation, Inc., Bethesda, Md., Feb. 1962. 256 p. [298] Jensen. R. A. The Function of the National Federa- tion of Science Abstracting and Indexing Services, Spec. Libraries 52, 555-557 (1961). [299] — A National Plan for Science Abstracting and Indexing Services, in H. P. Luhn [ed]. Automa- tion and Scientific Communication, Pt. 2, 1963, p. 339-342. [3001 — [ed]. Proceedings of the Annual Meeting 1962, National Federation of Science Abstracting and Indexing Services, Washington, D.C., 1963, 110 p. [301] — Proceedings of Sixth Annual Meeting. National Federation of Science Abstracting and Indexing Services, Washington, D.C.. 1963, 1 v. [302] A Joint College/Industry Library with Automata, prepared for Harvey Mudd College, Science and Engineering, Claremont. California, reprinted by the Council on Library Resources, Inc., Washing- ton, D.C., 1964. 35 p. [303] Jolley, L.. International ('.'inference on Cataloguing Principles TI. Thoughts After Paris, J. Docu- mentation 19, 47-62 (1963). [304] Juinpelt, R. W.. Towards a FIT-Policy in Scientific and Technical Translation, Babel 1, 21-25 (1955). [305] Kaiser, F. E., Translators and Translations: Serv- ices and Sources, Special Libraries Association, New York, 1959. [306] Kelsey, F. E., A National Drug Information Clear- inghouse, in R. A. Jensen [ed]. Proceedings, of Sixth Annual Meeting, 1963, p. 67-72. [307] Kenney, B. L., Cooperative Centralized Processing, A Report of the Establishment and First Year of Operation of the Southwest Missouri Library Service, Inc., American Library Association, Chicago, 1959, 98 p. [308] Kent, A., Cooperative Information Processing, in J. H. Shera et al. [eds], Documentation in Action, 1956, p. 27-41. [309] — Digital Computer Use in Medical Bibliography, Methods of Information in Medicine 1, 109-120 (1962). [310] — [ed], Information Retrieval and Machine Translation ( Part I ) , ( Based on the International Conference for Standards on a Common Language for Machine Searching and Translations, Cleve- land. Ohio. Sep 6-12. 1959), Interscience Publish- ers, New York, 1960, 686 p. [311] — Information Retrieval — Review and Prospectus, in C. M. Popplewell [ed], Information Processing 1962, 1963, p. 267-272. [312] — Textbook on Mechanized Information Retrieval, Interscience, New York. 1962. 268 p. [313] — Their Own Devices, Wilson Library Bull. 37, 276-277. 291 (1962). [314] Kent, F. L., Periodica Abbreviata and International Standardization, J. Documentation 10, 59-64 (1954). [315] Kilgour, F. G., R. T. Esterquist and T. P. Fleming, Computerization of Book Catalogues at the Columbia, Harvard and Yale Medical Libraries, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 2, 1963, p. 299-300. 131 [316] King, A., The Work of International Organizations and Foreign Documentation Centers, ASLIB Proc. 7,3-14 (1955). [317] King, E. F„ Electronic Transmission for Inter- library Loans, Amer. Documentation 11, 32-39 (1960). [318] King, G. W. [ed], Automation and the Library of Congress, (a survey sponsored by the Council on Library Resources, Inc.), Library of Congress, Washington, D.C., 1963, 88 p. [319] — The Automation of Library Systems, in B. E. Markuson [ed], Libraries and Automation, 1964, p. 233-244. [320] Knox, W. T., How Much Information Service, J. Chem. Documentation 2, 27 (1962). [321] Kochen, M., Toward Document Retrieval Theory : Relevance-Recall ratio for Texts Containing One Specified Query Term, in P. C. Janaske [ed], Automation and Scientific Communication, Pt. 3, 1964, p. 439-442. [322] Koelewijn, G. J., Recent Developments in Western Europe in the Field of the Automation of Docu- ment Retrieval Systems, Rev. Int. Doc. 29, 42-48 (1962). [323] Kraft, D. H., Application of IBM Equipment to Library Mechanization, Keyword-in-Context (KWIC) Indexing and the Selective Dissemina- tion of Information, in II. Goldhor [ed], Pro- ceedings of the 1963 Clinic on Library Applications of Data Processing, 1964, p. 133-156. [324] Kuentzel, L. E. [ed], Coden for Periodical Titles, ASTM Special Technical Publication No. 329, American Society for Testing and Materials, Philadelphia, 1963, 426 p. [325] Kyle, B. R. F., Information Retrieval and Subject Indexing : Cranfield and After, J. Documentation 20,55-69 (1964). [326] Ladd, D. L., ICIREPAT 1963— Progress, Pitfalls, and Prospects. (Press release of address for the third annual meeting of the Committee for Inter- national Cooperation in Information Retrieval among Examining Patent Offices (ICIREPAT) at the Patent Office of the Federal Republic of Austria, Vienna), Department of Commerce, Washington, D.C., 1963, 7 p. See also Keynote Address in H. Pfeffer [ed], Information Retrieval among Examining Patent Offices, 1964, p. 11-17. [327] Landau, T., [ed], Encyclopedia of Librarianship, Hafner Publishing Co., New York, 1958, 334 p. [328] Langenbeck, E. H., NOL Suggests Cooperative Cataloging. Plan to Reduce Information Costs Includes All Defensive Activites, Navy Manage- ment Review 6, 12-23, 17 (1962). [329] — A Plan to Reduce Costs of Technical Library Operations in the Department of Defense, Amer. Documentation 13, 295-300 (1962). [330] Larkey, S. V., The Army Medical Library Research Project at the Welch Medical Library, Bull. Med. Lib. Assoc. 37, 121-124 (1949). [331] — Cooperative Information Processing — Prospec- tus — Medicine, in J. H. Shera et al, [eds], Docu- mentation in Action, 1956, p. 301-306. [332] — The Welch Medical Library Indexing Project, Bull. Med. Lib. Assoc. 41, 32^0 (1953). [333] Lee, M. O., Responsibilities for Scientific Informa- tion in Biology: Proposal for Financing a Com- prehensive System, in National Academy of Sciences-National Research Council, Proceedings of the International Conference on Scientific In- formation, Vol. 2, 1959, p. 1417-1427. [334] Library Association, Gt. Britain, Scarborough Con- ference, Sep. 13-16, 1960. Proceedings, Papers and Summaries of Discussions, London 1960, 79 p. [335] — Southport Conference, Sep 19-23. 1955. Pro- ceedings, Papers and Summaries of Discussions, London, 1955, 101 p. [336] Library of Congress, Automation and the Library of Congress, Washington, D.C., 1963, 88 p. [337] — The Cataloging-in-Source Experiment, Wash- ington, D.C., 1960, 199 p. [338] — Subject Headings Used in the Dictionary Cat- alogs of the Library of Congress, 6th ed., Wash- ington, D.C., 1957, 1357 p. [339] — National Referral Center for Science and Tech- nology, A Directory of Information Resources in the United States, GPO, Washington, D.C., 1965, 352 p. [340] Liebesny, F., Co-operative Action: What Can Be Learned from Abroad, ASLIB Proc. 12, 149-151 (1960). Conference on Technical Translation, London, 15 January 1960. [341] — International Co-operation in Abstracting: An Experiment, in Association of Scientific Libraries and Information Bureaus, Looking Forward in Documentation, 1964, p. 4-19 to 4-21. [342] — Lost Information : Unpublished Conference Papers, in National Academy of Sciences-National Research Council, Proceedings of the Interna- tional Conference on Scientific Information, Vol. 1, 1959, p. 475^179. [343] Lipetz, B. A., Labor Costs, Conversion Costs, and Compatibility in Document Control Systems, Amer. Documentation 14, 117-122 (1963). [344] Little, Arthur D., Inc., Centralization and Docu- mentation, Final report to the National Science Foundation, C-64469, Cambridge, Mass., July 1963, 70 p., and Appendix, 1 v. [345] — Appendices to Final Report, Cambridge, Mass., July 1963, 178 p. [346] Lloyd, G. A., Select List of Standards and Other Aids for Editors and Publishers of Technical Literature, ASLIB Proc. 6, 282-286 (1954). [347] — Members' Forum : Standards in Documentation, ASLIB Proc. 7, 49-54 (1955). [348] Lockheed Aircraft Corp., An Evaluation of Infor- mation Retrieval Systems, Rept. no. 7170, Bur- bank, Cal., Sep. 30, 1959, 114 p. [349] Lubetzky, S., Code of Cataloging Rules: Author and Title Entry, American Library Association, Chicago, 1960, 86 p. [350] Luckett, G. R., Partial Library Automation with the Flexowriter Writing Machine, Lib. Res. Tech. Services 1, 207-210 (1957). [351] Ludington, F. B., The American Contribution to Foreign Library Establishment and Rehabilita- tion, Lib. Quarterly 24, 192-204 (1954). [352] Luhn, H. P., Auto-Encoding of Documents for In- formation Retrieval Systems, in M. Boaz [ed], Modern Trends in Documentation, 1959, p. 45-58. [353] — Automated Intelligence Systems, in L. H. Hat- tery and E. M. McCormick [eds], Information Retrieval Management, 1962, p. 92-100. [354] — [ed], Automation and Scientific Communica- tion, (Short Papers contributed to the Theme Sessions of the 26th Annual Meeting of the Amer- ican Documentation Institute, October 6-11, 1963 ) , American Documentation Institute, Wash- ington, D.C., 1963, Part 1, p. 1-128, Part 2, p. 129- 384. [355] — A Business Intelligence System, IBM J. Res. & Dev 2, 314-319 (1958). [356] — The IBM Electronic Information Searching Sys- tem, IBM Research Center, Yorktown Hights, N.Y., 1952, 1 v. [357] — Keyword-in-Context Index for Technical Litera- ture (KWIC Index), Rept. no. RC-127, IBM, Yorktown Heights, N.Y., Aug 31, 1959, 16 p. Also in Amer. Documentation 11, 288-295 (1960). [358] — A New Method of Recording and Searching In- formation, Amer. Documentation 4, 14-16 (1953). [359] — Selective Dissemination of New Scientific Infor- mation with the Aid of Electronic Processing Equipment, IBM Research Center, Yorktown Heigths, N.Y., Nov 30, 1959, 19 p. Also in Amer. Documentation 12, 131-138 (1961). 132 £360] — A Statistical Appraoeh to Mechanized Encoding and Searching of Literary Information, IBM J. Res. & Dev. 1, 309-317 (1957) . {361] Lundy, J. T., Representation of Complex Character Sequences on an Expandable Input Keyboard, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 1, 1963, p. 115-116. £362] Lunin, L. F. and M. Stovall, Theasurus for Litera- ture Analysis in a Cancer Research Institution : Construction and Correlation with Three National Thesauri, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 2, 1963. £363] Maass, W. G., New Information Services from a Not-So-Old Publishing House, J. Chem. Docu- mentation 2, 46-48 (1962). £364] MacDonald, M. R., Cataloging at the Armed Forces Medical Library, 1945-1952, J. Cataloging and Classification IX, 58-78 (1953). £365] MacPherson, J., World Communications : The Im- pact of Speed, Nature 168, 446-448 (1951). £366] MacWatt, J. A, Improving Scientific Communica- tion : Reprints Directly Available From The Publisher at a Reasonable Fee Could Supplement Today's Journals, Science 134, 313-316 (1961). £367] Mahany, P. R., W. A. Creager and S. Herner, Functional Symbols Currently in Use in Gov- ernment Reports, Final report on Contract NSF- 213, Herner and Co., Washington, D.C., Dec 1962, 121 p. [368] Maloney, C. J., Abstract Theory of Retrieval Cod- ing, in National Academy of Sciences-National Research Council, Proceedings of the Interna- tional Conference on Scientific Information, Vol 2, 1959, p. 1365-1282. £369] Maloney, C. J. and H. W. Batchelor, Tentative Principles of Information Retrieval in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 1, 1963, p. 107-108. £370] Marden, E. and H. R. Roller, A Survey of Com- puter Programs for Chemical Information Searching, NBS Technical Note 85, Office of Tech- nical Services, Washington, D.C., Feb 1961, 84 p. £371] Markus, J., State of the Art of Published Indexes, Amer. Documentation 13, 15-30 (1962). £372] Markuson, B. E., [ed], Libraries and Automation, (Proceedings of the Conference on Libraries and Automation held at Airlie Foundation, Warren- ton, Virginia, May 26-30, 1963, under sponsorship of the Library of Congress, National Science Foundation, Council of Library Resources, Inc.), Library of Congress, Washington, D.C., 1964, 268 p. [373] Martin, B., JPRS : Largescale Government Trans- lation Service, Linguistic Reporter 4, 1-2 (1962). [374] Masterman, M., Potentialities of a Mechanical Thesaurus, Mech. Translation 3, 36 (1956). (Paper presented at the Second International Conference on Machine Translation, Held at the Massachusetts Institute of Technology, Cam- bridge, Mass., on Oct 16-20, 1956). £375] — The Thesaurus in Syntax and Semantics, Mech. Translation 4, 35-43 (1957). [376] Mattson, L., UNESCO Science Cooperation Office for Southeast Asia and its Activities, Amer. Documentation 13, 328-333 (1962). £377] McGee, L. L., W. J. Holliman, Jr., A. Z. Loren and G. D. Adams, Compilation and Computer Up- dating of a Medical Sciences Thesaurus, in H. P. Luhn [ed], Automation and Scientific Communi- cation, Pt. 2, 1963. p. 347-348. £378] Menzel, H., Planned and Unplanned Scientific Com- munication, In National Academy of Sciences- National Research Council, Proceedings of the In- ternational Conference on Scientific Information, Vol 1, 1959, p. 199-243. £379] Mersel, J. and S. B. Smith, Center for Text in Machine-Usable Form, a feasibility study under contract NSF-C 320, TRW Computer Division, Thompson Ramo Wooldridge, Inc., Canoga Park, Cal., Feb. 28, 1964, 103 p. [380] Metcalf, K. D., J. Doe, T. P. Fleming, M. L. Mar- shall, L. Q. Mumford and A. D. Osborn, The Na- tional Medical Library Report of a Survey of the Army Medical Library, American Library Asso- ciation, Chicago, 1944, 94 p. [381] Metcalfe, J. W., Information Indexing and Subject Cataloging: Alphabetical, Classified, Coordin- ate, Mechanical, Scarecrow Press, New York, 1957, 338 p. [382] — Subject Classifying and Indexing of Libraries and Literature, Scarecrow Press, New York, 1959, 347 p. [383] Microfiche Standards Adopted, Chem. & Eng. News 42,54-55 (1964). [384] Mikhailov, A. I., On the Functioning of the All- Union Institute for Scientific and Technical Infor- mation of the USSR Academy of Sciences, in Na- tional Academy of Sciences-National Research Council, Proceedings of the International Confer- ence on Scientific Information, Vol. 1, 1959, p. 511-521. [385] — Problems of Mechanization and Automation of Information Work, Rev. Int. Doc. 29, 49-56 (1962). [386] Miller, E. E., The Genesis and Characteristics of Report Literature, Amer. Documentation 3, 91-94 (1952). [387] Miller, T., Six Minneapolis "Insiders" Build Unique Cooperative, Spec. Libraries 54, 295-297 (1963). [388] Mills, J., The Universal Decimal Classification, Vol. I of the Rutgers series on Systems for the Intellectual Organization of Information. S. Artandi [ed], Graduate School of Library Science, Rutgers University, New Brunswick, N. J., 1964, 132 p. [389] Minder, T. and G. Lazorick, Automation of the Penn State University Acquisitions Department, in P. C. Janaske [ed], Automation and Scientific Communication, Pt. 3, 1964, p. 455-459. [390] Moffit, A., Punched Card Records in Serials Ac- quisition, College and Research Libraries 7, 10-13 (1946). [391] Mohrhardt, D. E., Statement, in U.S. Congress, House Committee on Science and Astronautics, Dissemination of Scientific Information, 1959, p. 27-40. [392] Mooers, C. N., The Indexing Language of an Infor- mation Retrieval System, in W. Simonton [ed]. Information Retrieval Today, 1963, p. 21-36. [393] — The Tape Typewriter Plan: a Method for Co- operation in Documentation. Rept. no. ZTB-137, Contract AF49 (638) 376, Zator Co., Cambridge, Mass., July 1960, 22 p. Also in ASLIB Proc. 12, 277-291 (1960). [394] — The Zator — A Proposal, A Machine Method for Complete Documentation. Technical Bulletin no. 65, Zator Co., Boston. Mass., 1951, 20 p. [395] Morris. J. C, The Duality Concept in Subject Analysis, Amer. Documentation 5, 117-146 (1954). [396] — Evolution or Involution? Notes Critical of the Uniterm System of Indexing, J. Cataloging and Classification 10, 111-121 (1954). [397] Morris, M. E. L., From One to Two Fifty: The Economics of Low Cost Bookshelf Replenishment In Association of Special Libraries and Informa- tion Bureaus, Looking Forward in Documenta- tion, 1964, p. 3-1 to 3-4. [398] Morse, R. D., A. I. Ch. E. Information-Retrieval Activities, Amer. Documentation 13, 69-70 (1962). [399] — Information Retrieval, Chem. Eng. Progress 57, 55-58 (1961). Also in Engineers Joint Council, Proceedings of the EJC Information Svmposium, 1962, p. 81-84. 133 [400] Mumford, L. Q., Statement, in U.S. Congress, House Committee on Science and Astronautics, Dissemination of Scientific Information, 1959, p. 1-16. [401] Murra, K. O., History of Some Attempts to Or- [420] ganize Bibliography Internationally, in J. H. Shera and M. E. Egan [eds], Bibliographic Or- ganization, 1901, p. 24-53. [421] [402] — [ed], International Scientific Organizations: a Guide to their Library, Documentation, and In- [422] formation Services, Library of Congress, Wash- ington, D.C., 1962, 794 p. [403] National Academy of Sciences-National Research [423] Council, Communication Problems in Biomedical Research, Report of a Study by the Division of [424] Medical Sciences National Academy of Sciences- National Research Council in cooperation with the Federation of American Societies for Experi- mental Biology and with the Institute for Ad- [425] vancement of Medical Communication, Washing- ton, D.C., Oct 31, 1963, 24 p. [404] — Proceedings of the International Conference on Scientific Information, (held at Washington, [426] D.C., November 16-21, 1958), Washington, D.C., 1959, Vol. 1, p. 1-813, Vol. 2, p. 817-1635. [405] National Aeronautics and Space Administration, Scientific and Technical Aerospace Reports, Cumulative Indexes, Jan-Mar 1963, 1 v. [406] National Agricultural Library, Subject Heading [427] List, Washington, D.C., July 1963, 4 v. [407] National Bureau of Standards, National Standard Reference Data Program : Background Informa- [428] tion, XBS Technical Note 194, Washington, D.C., June 1963, 18 p. [408] National Federation of Science Abstracting and In- dexing Services, A Guide to U.S. Indexing and Abstracting Services in Science and Technology, [428] NFSAIS rept. no. 101, Washington, D.C., 1960, 79 p. [409] — A Guide to the World's Abstracting and Index- ing Services in Science and Technology, NFSAIS rept. no. 102, Washington, D.C., 1963, 183 p. [429] [410] National Federation of Science Abstracting and Indexing Services, Transliteration of the Modern Russian Alphabet, NFSAIS Standard no. 63-1, [430] Washington, D. C. 1963. 1 p. [411] National Library of Medicine, Medical Subject Headings: Main Headings and Cross References Used in Index Medicus and National Library of [431] Medicine Catalog, 2nd ed., Part 2 of Index Medi- cus. Vol. 4. no. 1. Jan. 1963, 317 p. [412] — The MEDLARS Story at the National Library of Medicine, Washington. D. C, 1963, 74 p. [432] [413] — The National Library of Medicine Index Mecha- nization Project, Part 2 of Bull. Med. Lib. Assoc, Vol. 49, no. 1, Jan. 1961, 96 p. [433] [414] National Microfilm Association, Basic Standards and Specifications of the Department of Defense Engineering Data Microreproduction System, [434] NMA informational monograph No. 1, Annapolis, Md., 1963, 151 p. [415] National Resources Committee, Research — a National Resource. I. Relation of the Federal Government to Research, U.S. Government [435] Printing Office, Washington, D.C., 1938, 255 p. [416] National Resources Planning Board, Research — a National Resource. II. Industrial Research, U.S. Government Printing Office, Washington, [436] D.C., 1941, 370 p. [417] The National Science Foundation . . ., Library of Congress Inf. Bull. 22, 545-546 (1963). [418] National Science Foundation. Current Research and [437] Development in Scientific Documentation, No. 11, NSF-63-5, Office of Technical Services, Washing- ton, D.C., Nov. 1962, 440 p. [438] [419] — • Specialized Science Information Services in the United States. A Directory of Selected Special- ized Information Services in the Physical and, Biological Sciences, NSF-61-68, U.S. Government Printing Office, Washington, D.C., Nov. 1961, 528 p. The Need for a Faceted Classification as the Basis of All Methods of Information Retrieval, Library Assoc. Record 57, 262-268 (1955). Neiswender, R., Russian Transliteration — Sound and Sense, Spec. Libraries 54, 37-11 (1962). New ASA Standard Offers Coding Uniformity for Data Interchange, IBM Data Processor 7, Aug. 7, 1963, 12 p., special issue. New Tools for the Resurrection of Knowledge, Chem. & Eng. News 32, 866-869 (1954). Newman, S. M., Classified Patent Search Files, a Proposed Base for Technical Information Centers, Amer. Documentation 12, 206-209 (1961). Also in J. Pat. Off. Society 43, 418-434 (1961). — Information Retrieval Research in the U.S. Patent Office, MULL (Modern Uses of Logic in Law) 1, 45-55 (1960). Also in J. Pat. Off. Society 42, 731-741 (1960). Newman, S. M., Storage and Retrieval of Contents of Technical Literature : Nonchemical Informa- tion. First Supplementary Report, Patent Office Research and Development rept. No. 4, Washing- ton, D.C., 1957, 16 p. See Andrews, D. D. and S. M. Newman for original report. — ■ Second Supplementary Report, Patent Office Re- search and Development rept. No. 12, Washington, D.C., 1958, 12 p. — R. W. Swanson and K. C. Knowlton, A Notation System for Transliterating Technical and Scien- tific Texts for Use in Data Processing Systems, in A. Kent [ed], Information Retrieval and Machine Translation (Part I), 1960, p. 345-376. Newman, S. M., R. W. Swanson and K. C. Knowlton, A Notation System for Transliterating Technical and Scientific Texts for Use in Data Processing Systems, in A. Kent [ed], Information Retrieval and Machine Translation (Part I), 1960, p. 345- 376. Niwa, Y. X., Present State of Mechanization of Doc- umentation in Japan, Rev. Int. Doc. 29, 63-65- (1962). Nugent, W. R., The Automated Editing of Text via Computer and Off-Line Devices, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 1, 1963, p. 125-126. — ■ A Machine Language for Document Translitera- tion, preprint, 14th National Conference of the Association for Computing Machinery, Cambridge, Mass.. Sep 1959, 33 p. Oak Ridge National Laboratory, Key Word Index : Laboratory Reports Received, Semiannual index, Jan-June 1963, Oak Ridge, Tenn., 89 p. O'Brien, I. C, Recent Activities in Foreign Ex- changes, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 2, 1963, p. 203-204. O'Donnell, T. R., J. L. Lewis and J. I. Glendinning, Federal Government's System for Distributing Its Unclassified R&D Reports. Contract NSF C-224, John I. Thompson Co., Washington, D.C., June 20, 1962. 1 v. Oellrich, G. L., Cataloging and Classification, in H. Orlans [ed], Federal Departmental Libraries: a Summary Report of a Survey and a Conference, 1963, p. 67-72. Orlans, H., [ed], Federal Departmental Libraries: A Summary Report of a Survey and a Conference, Brookings Institute, Washington, D.C., 1963, 150 p. Orne, J., Resources of Foreign Scientific Literature : Acquisition on a National Scale, Amer. Documen- tation 14, 229-233 (1963). Ornsen, K. C. and R. L. Paddock, Industry-Commit- tee Guidance of a Centralized Abstracting Service, J. Chem. Documentation 1, 45-47 (1961). 134 [439] Orr, R. H., Letter to the Editor, Amer. Documenta- tion 14, 330-331 (1963) . [440] Orr, R. H., G. Abdian, C. P. Bourne, E. B. Coyl, A. A. Leeds, and V. M. Pings, The Biomedical Information Complex Viewed as a System, Fed- eration Proceedings 23, 1133-1145 (1964). [441] Orr, R. H. and V. M. Pings, Document Retrieval : The National Biomedical Library System and Interlibrary Loans, Federation Proceedings 23, 1155-1163 (1964). [442] Orr, R. H., V. M. Pings and A. A. Leeds, Reference Retrieval Tools : Biomedical Abstracting and In- dexing Services, Federation Proceedings 23, 1164- 1176 (1964). [443] Osborn, A. D., Cataloging and Cataloging Codes in Other Countries Today, Lib. Quarterly 26, 276- 285 (1956). [444] — The Future of the Union List of Serials. College and Research Libraries 15, 26-28, 118 (1954). [445] Painter, A. F., An Analysis of Duplication and Consistency of Subject Indexing Involved in Re- port Handling at the Office of Technical Services, Contract CC-4753, Office of Technical Services, Washington, D.C., Mar. 1963, 135 p. [446] Parker-Rhodes, A. F., Mechanical Translation Pro- gram Utilizing an Interlingual Thesaurus, Mech. Translation 3, 36 ( 1956 ) . ( Paper presented at the Second International Conference on Machine Translation, held at the Massachusetts Institute of Technology, Cambridge, Mass., on Oct. 16-20, 1956). [447] Parkins, P. V., New Section on Communication, Science 135, 535-536 (1962). [448] Parthasarathy, S.. INSDOC (Indian National Scien- tific Documentation Centre) and Its Regional Activities, Amer. Documentation 13, 334-337 (1962). [449] Patrick, L. G., Experience of Co-operation in Trans- lation Work, ASLIB Proc. 12, 131-135 (1960). Conference on Technical Translation, London, 15 Jan. 1960. [450] Patrick, R. L. and D. V. Black, Index Files: Their Loading and Organization for Use, in B. E. Markuson [ed], Libraries and Automation. 1964, p. 29-53. [451] Pearl, R., Some Contributions of Dr. John Shaw Billings to the Development of Vital Statistics, Bull. Inst. History of Medicine 6, 387-393 (1938). [452] Peiss, R.. Report on Europe. College and Research Libraries 8, 113-119 (1947). [453] Perez-Vitoria. A.. Document UNESCO NX 177— a Code of Good Practice for Scientific Publications, Amer. Documentation 14, 241-243 (1963). [454] Perry, J. W., Cooperative and Centralized Informa- tion Processing — Dream or Reality?, in J. H. Shera et al. [eds], Documentation in Action, 1956, p. 27-41. [455] Perry, J. W., A. Kent and M. M. Berry, Machine Literature Searching, Interscience, New York, 1956, 162 p. [456] Pfeffer, H., [ed], Information Retrieval Among Examining Patent Offices, (Third Annual Meet- ing of the Committee for International Coopera- tion Among Examining Patent Offices (ICIREPAT) at the Patent Office of the Federal Republic of Austria, Vienna, September 17th through 26th, 1963), Spartan Books, Baltimore, Md., 1964, 347 p. [457] Phelps, R. H. and J. P. Herlin, Alternatives to the Scientific Periodical : a Report and a Bibliog- raphy, UNESCO Bull. Libraries 14, 61-75 (1960). [458] Pings. V. M., Universal Classification— Untenable, Library Res. Tech. Services 4, 5-13 (1960). [459] Poindron, P., Availability of Scientific Conference Papers and Proceedings, UNESCO Bull. Li- braries 16, 165-176 (1962). [460] — La Conference Internationale sur les Principes de Catalogage, Rev. Int. Doc. 29, 18-22 (1962). [461] — ■ The Content, Influence and Value of Scientific Conference Papers and Proceedings, UNESCO Bull. Libraries 16, 113-126 (1962). [462] Popplewell, C. M., [ed], Information Processing 1962, (Proceedings of IFIP Congress, Munich, Aug 27-Sep. 1, 1962), North-Holland Publishing Co., Amsterdam, 1963, 780 p. [463] President's Science Advisory Committee, Improv- ing the Availability of Scientific and Technical Information in the United States, White House Press Release for Dec 7, 1958, 9 p. [464] — Science, Government, and Information, (Wein- berg report) U.S. Government Printing Office, Washington, D.C., 1963, 52 p. Also in U.S. Con- gress, House Committee on Education and Labor, National Information Center, Hearings ... on H.R. 1946, Appendix to Vol 1, 1963, p. 453-508. [465] — Scientific and Technological Communication in the Government, (Crawford Task Force on Fed- eral Scientific and Technological Information Problems), Washington, D.C., Apr 1962, 81 p. Also in U.S. Congress, House Committee on Edu- cation and Labor, National Information Center, Hearings ... on H.R. 1946, Appendix to Vol 1, 1963, p. 21-107. [466] Proceedings of the Librarians' Convention held in New York City, September 15, 16, and 17, 1853, Reprinted for William H. Murray, The Torch Press, Cedar Rapids, Iowa, 1915, 63 p. [467] Randall, G. E., Man is Measured by His Horizons, Spec. Libraries 53, 380-381 (1962). [468] Ranganathan, S. R., Colon Classification and Its Approach to Documentation, in J. H. Shaer and M. E. Egan [eds], Bibliographic Organization, 1951, p. 94-105. [469] — Prolegomena to Library Classification, 2nd ed., Library Association, London, 1957. [470] — Whv "Documentation"?, ASLIB Proc. 4, 105-108 1952). [471] Ray, L. C, Keypunching Instructions for Total Text Input, in American University, Machine In- dexing : Progress and Problems, 1962, p. 50-57. [472] Readett, A. G.. The Benefits of Co-operation, ASLIB Proc. 12, 152-157 (1960). Conference on Technical Translations, London, 15 January 1960. [473] Recent Actions of the FCST Committee on Scientific Information, Scientific Information Notes 5, 2-3 (1963). [474] Redman, H. F. and L. E. Godfrey [eds]. Dictionary of Report Series Codes, Special Libraries Asso- ciation, New York, 1962, 648 p. [475] Reed, J. B., Some Impediments to Co-operation in Translations : Copyright and Other Matters, ASLIB Proc. 12, 158-166 (1960). Conference on Technical Translation. London, 15 January 1960. [476] Regional Technical Report Centers Established, Scientific Information Notes 4, 1-2 (1962). [477] Reiteration of ACM Policy Toward Standardization, Comm. Assoc. Computing Machinery 5, 547-549 (1962). [478] Reynolds, W. L., The Senate Committee on Govern- ment Operations and Documentation, Amer. Documentation 12, 93-97 ( 1961 ) . [479] Rhees, W. J. [ed]. The Smithsonian Institution: Documents Relative to Its Origin and History, 1835-1899, Smithsonian Institution Publications 1035 and 1036, U.S. Government Printing Office. Washington, D.C., 1901, 2 v. 1938 p. [480] Richardson, E. C, Some Aspects of International Library Cooperation, F. S. Cook & Son, Yardley, Pa., 1928, 168 P. [481] Richardson, H. G., The Proposed Houston Technical Information Center, Spec. Libraries 54, 297-299 (1963). [482] Richmond, P. A., Review of the Cranfield Project, Amer. Documentation 14, 307-311 (1963). 135 [483] Richter, A. J., Z-39 Today: Its Work and Its Subcommittees, Spec. Libraries 54, 107-109 (1963). [484] Ridenour, L. N., R. R. Shaw and A. G. Hills, Bibli- ography in An Age of Science, Univ. of Illinois Press, Urbana, 1951, 90 p. [485] Rider, F., The Possible Correlation of All Forms of Microtext, Amer. Documentation 2, 152-157 (1951). [486] — The Scholar and the Future of the Research Library : a Problem and Its Solution, Hadham Press, New York, 1944, 236 p. [487] Roberts, M. A., The Library of Congress in Relation to Research, in National Resources Committee, Research-A National Resource. I. Relation of the Federal Government to Research, 1938, p. 235-255. [488] Rogers, F. B., Facilities and Services of the Na- tional Library of Medicine, Bull. Millard Fillmore Hospital, Buffalo, N.Y., 7, 32-38 (1960). [489] — Stresses in Current Medical Bibliography, New England J. Medicine 267, 704-708 (1962). [490] — and S. Adams, The Army Medical Library's Publication Program, Texas Reports on Biology and Medicine 8, 271-300 (1950). [491] Rolling, L., A Keyword Index for Machine Docu- mentation in the Nuclear Field, EUR/CID/4243/ 10/62/e, EURATOM, Brussels, Aug. 1962, 1 v. [492] Rolling, L, Un Repertoire de Mots-Cles pour la Documentation Mechanisee dans le Domaine de la Technique Nucleaire, Bull. Bibliotheques de France 8, 11-25 (1963). [493] Ross, H. McG., Considerations in Choosing a Char- acter Code for Computers and Punched Tapes, The Computer Journal 3, 202-210 (1961). [494] — Current Position on Standards Work Relating to Computers, The Computer Bulletin 6, 133-135 (1963). [495] Rossheim, R. J., Report on Proposed American Standard Flowchart Symbols for Information Processing, Comm. Assoc, of Computing Ma- chinery 6, 599-604 (1963). [496] Rothgeb, C. L., Centralized Processing of Technical Information at the Scientific and Technical In- formation Facility, in H. P. Luhn [ed], Automa- tion and Scientific Communication, Pt. 2, 1963 p. 277-278. [497] Rothstein, S., The Development of the Concept of Reference Service in American Libraries, 1850-1900, Lib. Quarterly 23, 1-15 (1953). [498] Royal Society of London, Royal Society Scientific Information Conference, London, June 21-July 2, 1948 (Report and Papers Submitted), London, 1948, 723 p. [499] Rubenstein, A. H., Timing and Form of Research- ers' Needs for Techical Information, J. Chem. Documentation 2, 28-31 (1961). [500] Rubinoff, M.. Data Retrieval, text of a presentation at the Technical Forum of the Insitute of Elec- trical and Electronics Engineers, Toronto, June 18, 1963, 4 p. [501] Runge, G. E., Ted], Correlation Index: Document Series and PB Reports, Special Libraries Associ- ation, New York, 1953, 271 p. [502] Salton, G., [ed], Information Storage and Retrieval, Scientific rept. no. ISR-3, Computation Labora- tory, Harvard University, Cambridge, Mass., 1 Apr 1963, 1 v. [503] Sammet, J. E., H. Ohlman and H. G. Bohnert, Discussing Pros and Cons of a Special IR Lan- guage, Comm. Assoc. Computing Machinery 5, 8-10 (1962). [504] Savage, T. R., Is Yehoshua Bar-Hillel Approaching a Crisis?, Amer. Documentation 14, 331 (1963). [505] Schuchmann, M., Bibliography of Standards on Documentation, FID publication no, 303, Inter- national Federation of Documentation, The Hague, 1961, 107 p. ; supplement, 1961, 11 p. [506] Schuller, J. A., Experience with Indexing and Re- trieving by UDC and Uniterms, ASLIB Proc. 12, 372-389 (1960). [507] Schullian, D. M. and F. B. Rogers, The National Library of Medicine, Lib. Quarterly 28, 1-17, 95-121 (1958). [508] Schultheiss, L. A., D. S. Culbertson and E. M. Heiliger, Advanced Data Processing in the Uni- versity Library, Scarecrow Press, New York, 1962, 388 p. [509] Schultz, C. K, Editing Author-Produced Indexing Terms and Phrases Via a Magnetic-Tape Thesau- rus and a Computer Program, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 1, 1963, p. 9. [510] — An Imaginary Panel Discussion About Indexing, in American Documentation Institute, Param- eters of Information Science, Vol. 1, 1964, p. 437- 452. [511] — and C. A. Schwartz, A Generalized Computer Method for Index Productions, Amer. Documen- tation 13, 420-432 (1962). [512] — and C. A. Shepherd, The 1960 Federation Meet- ing: Scheduling a Meeting and Preparing an Index by Computer, Federation Proceedings 19, 682-699 (1960). [513] Schwegmann, G. A., Jr., The Bibliographical Control of Microforms, Library Trends 8, 380-390 (1960). [514] Sectional Committee Z39, J. Chem. Documentation 3,123 (1963). [515] Senders, J. W., Information Storage Requirements for the Contents of the World's Libraries, Science 141,1067-1068 (1963). [516] Sewell, W., Cooperative Information Processing — Pharmaceuticals (Summary), in J. H. Shera et al [eds], Documentation in Action, 1956, p. 267-268. [517] Shachtman, B. E., Documents and Information Exchange, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 2, 1963, p. 257-258. [518] Shaw, R. R.. Documentation and the Individual, Science 137, 409-411 (1962). [519] — Information Retrieval, Science 140, 606-609 (1963). [520] — Parameters for Machine Handling of Alphabetic Information, Amer. Documentation 13, 267-269 (1962). [521] — The Publication Board, College and Research Libraries 7,105-108 (1946). [522] — [ed], The State of the Library Art, Rutgers Uni- versity Press, New Brunswick, N.J., 1960-61, 5 v. [523] Shera, J. H., Classification as the Basis of Biblio- graphic Organization, in J. H. Shera and M. E. Egan [eds], Bibliographic Organization, 1951, p. 72-93. [524] — and M. E. Egan [eds], Bibliographic Organiza- tion (Papers presented before the Fifteenth An- nual Conference of the Graduate Library School July 24-29, 1950), The Univ. of Chicago Press, Chicago, 111., 1951, 275 p. [525] — A. Kent and J. W. Perry [eds], Documentation in Action, (Proceedings of Conference on the Practical Utilization of Recorded Knowledge Present and Future), Reinhold Publishing Corp., New York, 1956, 471 p. [526] — Information Resources : A Challenge to Amer- ican Science and Industry, (Based upon the Pro- ceedings of a Special Meeting of the Council on Documentation Research. February 3-4, 1958, Western Reserve University), Western Reserve University Press, Cleveland, Ohio, 1958, 214 p. [527] Sherrod, J., Feasibility Criteria for Establishing Specialized Information Centers, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 2, 1963, p. 221. 136 [528] — Functions of a Technical Information Center, [547] in L. H. Hattery and E. M. McCormick [eds], Information Retrieval Management, 1962, p. 34- 36. [548] [529] — A Progress Report on an Experiment in Semi- automatic Indexing Conducted by the AEC Divi- [549] sion of Technical Information Extension, in H. P. Luhn [ed], Automation and Scientific Communi- cation, Pt. 2, 1963, p. 215. [550] [530] Shilling, C. W., Requirements for a Scientific Mis- sion-Oriented Information Center, Amer. Docu- mentation 14, 49-53 (1963). [531] — Status Report on the Biological Sciences Com- [551] muncation Project (BSCP), in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 2, [552] 1963, p. 205-206. [532] Shipman, F. W., Weeding, in H. Orlans [ed], Fed- eral Departmental Libraries : A Summary Report of a Survey and a Conference, 1963, p. 62-66. [553] [533] Simonton, W. [ed], Information Retrieval Today (Papers presented at the Institute, conducted by the Library School and the Center for Continua- [554] tion Study, University of Minnesota, September 19-22, 1962), University of Minnesota, 1963, 176 p. [555] [534] Slamecka, V., Classificatory, Alphabetical, and Associative Schedules as Aids in Coordinate In- dexing, Amer. Documentation 14, 223-228 (1963). [535] — and J. Jacoby, Effect of Indexing Aids on the Re- [556] liability of Indexers, Final Technical note, Con- tract AF30( 602) -2616, Documentation, Inc., [557] Bethesda, Md., June 1963, 1 v. [536] Smith, A., Substantive Changes in the Draft Code, June 1960-July 1961, Library Res. Tech. Services 5,341-342(1961). [558] [537] Smith, W. A., Locating Periodicals. I. The World List of Scientific Periodicals, Library Assoc. Record 55, 245-247 (1953) . [538] Smithsonian Institution, Report on the Interna- [559] tional Exchange Service for the Year Ended June 30, 1963, U.S. Government Printing Offlce, Washington, D.C., 1964, p. 64-73 from the Smithsonian Report for 1963. [° 6 °] [539] Special Librarians and the Weinberg Report, Spec. Libraries 54, 325-332 (1963). [561] [540] Special Libraries Association, Guide to the SLA Collection of Classification Schemes and Subject Heading Lists on Deposit at Western Reserve [562] University as of March 20, 1961. B. R. Barden and B. Denison [comp], New York, 1961, 104 p. [541] Sprenkle, P. M. and F. G. Kilgour, A Quantitative Study of Characters on Biomedical Catalogue [563] Cards — A Preliminary Investigation, Amer. Documentation 14, 202-205 (163). [542] Stanford Research Institute, A Draft Program for [564] a National Technical Information Center Pre- pared by Stanford Research Institute, in U.S. Congress, House Committee on Education and [565] Labor, National Information Center. Hear- ings ... on H.R. 1946, Appendix to Vol. 1, 1963, p. 1-20. [543] Stevens, M. E., Availability of Machine-Usable [° 66 1 Natural Language Material, in American Uni- versity, Machine Indexing Problems and Progress, 1962, p. 58-75. [567] [544] — and E. De Grolier, Summary of Discussion. Area 4, Organization of Information for Storage [568] Search. Comparative Characteristics of Exist- ing Systems, in National Academy of Sciences- [569] National Research Council, Proceedings of the International Conference on Scientific Informa- tion, 1959, Vol. 1, p. 803-812. [545] Stiles, H. E., The Association Factor in Information [570] Retrieval, J. Assoc. Computing Machinery 8, 271- 279 (1961). [546] Strieby, I. M., Cooperative Information Processing [571] — Pharmaceuticals, in J. H. Shera et al [eds], Documentation in Action, 1956, p. 344-347. Structures of Standards-Processing Organizations in the Computer Area, Comm. Assoc. Computing Machinery 6, 294-305 (1963). A Subcommittee on Classification, Library of Con- gress Inf. Bull. 22, 630 (1963). Swanson, D. R., Design Requirements for a Future Library, in B. E. Markuson [ed], Libraries and Automation, 1964, p. 11-21. — Information Retrieval ; State of the Art, in Pro- ceedings of the American Federation of Informa- tion Processing Societies, Joint Computer Conference 1961, p. 239-246. — Library Goals and the Role of Automation, Spec. Libraries 53, 466-471 (1962). Taine, S. I., Bibliographic Data Processing at the National Library of Medicine, in H. Goldhor [ed], Proceedings of the 1963 Clinic on Library Applica- tions of Data Processing, 1964, p. 109-132. — The Future of the Published Index, in American University, Machine Indexing Problems and Prog- ress, 1962, p. 144-149. Talmadge, R. L., The Farmington Plan Survey : an Interim Report, Coll. Res. Libraries 19, 375-383 (1958). Tareev, B. M., Methods of Disseminating Scientific Information, and Science Information Activities in the USSR, Amer. Documentation 13, 338-343 (1962). Taube, M., The Cataloging of Publications of Cor- porate Authors, Lib. Quarterly 20, 1-20 (1950). — Experiments With the IBM-9900 and a Discus- sion of an Improved Comae as Suggested by These Experiments, J. Chem. Documentation 2, 22-76 (1962). — Extensive Relations as the Necessary Condition for the Significance of Thesauri for Mechanized Indexing, J. Chem. Documentation 3, 177-180 (1963). — Functional Approach to Bibliographic Organiza- tion : a Critique and a Proposal, in J. H. Shera and M. E. Egan [eds], Bibliographic Organization, 1951, p. 57-71. — Possibilities for Cooperative Work in Subject Controls, Amer. Documentation 3, 21-28 (1952). — Storage and Retrieval of Information by Means of the Association of Ideas, Amer. Documenta- tion 6, 1-18 (1955). Taube, M. and Associates, Studies in Coordinate Indexing, Documentation, Inc., Washington, D.C., Vol. 1, 1953; Vol. 2, 1954; Vol. 3, 1956; Vol. 4, 1957 ; Vol. 5, 1959. Tauber, M. F., Cataloging and Classification, in R. R. Shaw [ed], The State of the Library Art, Vol. 1, part 1, 1960, p. 167-251. Tauber, M. F. and Associates, Technical Services in Libraries, Columbia University Press, New York, 1954, p. Taylor, K. L. Reader Services, in H. Orlans [ed], Federal Departmental Libraries : A Summary Re- port of a Survey and a Conference, 1963, p. 73-77. Technical Translation : Co-operative Action, ASLIB Proc. 12, 129-168 (1960). ASLIB Conference, London, 15 January 1960. Technical Reparations, Federal Science Progress 1, no. 3, 8-9 (1947). "There'll Be More Microfilming Soon North of the Border", Systems Management 3, 47-48 (1962). Thompson, E., [ed], A.L.A. Glossary of Library Terms with a Selection of Terms in Related Fields American Library Association, Chicago, 1943, 159 p. Thompson, M., Automatic Reference Analysis, in G. Salton [ed], Information Storage and Retrieval, Scientific rept. no. ISR-3, 1963, p. II-1-II-27. Towner, I. L., [comp.], Directory of Special Librar- ies, Special Libraries Association, New York, 1953, 297 p. 137 [572] True, W. P., The First Hundred Years of the Smith- sonian Institution 1846-1946, [Smithsonian In- stitution], Washington, D.C., Aug. 10, 1946, 46 p. [573] Union List of Serials in Libraries of the United States and Canada, W. Gregory [ed], 2nd ed., H. W. Wilson, New York, 1943, 3063 p. [574] — Supplement to the 2nd ed., Jan. 1941-Dec. 1943, G. E. Malikoff [ed], H. W. Wilson, New York, 1945, 1123 p. [575] — Second Supplement to the 2nd ed., Jan. 1944- Dec. 1949, M. Franck [ed], H. W. Wilson, New York, 1953, 1365 p. [576] United Nations Educational, Scientific and Cul- tural Organizations, Information Processing, (Proceedings of the International Conference, Paris, 15-20 June 1959), Oldenbourg, Munich; Butterworths, London, 1960, 520 p. [577] U.S. Air Force, Air Force Technical Objective Document [TOD] Release Program, 1964, 7 p. [578] U.S. Atomic Energy Commission, The Literature of Nuclear Science : Its Management and Use, (Proceedings of conference held at Division of Technical Information Extension, Oak Ridge, Tenn., Sep. 11-13, 1962), TID-7647, Oak Ridge. Tenn., 1962, 398 p. [579] — Research and Development Abstracts of the USAEC, RDA-3, Oak Ridge, Tenn., July-Sep 1962, and RDA-4, Oct.-Dec. 1962. Issues contain an experiment in automatic indexing). [580] — Subject Headings Used by the USAEC Techni- cal Information Service, TID-5001, 3rd ed., Oak Ridge, Tenn., 1960, 476 p. [581] — Subject Headings Used by the USAEC Division of Technical Information, TID-5001, 4th ed., Oak Ridge, Tenn., 1962, 375 p. [582] U.S. Congress, House Committee on Education and Labor, National Information Center. Hearings ... on H.R. 1946, 88th Cong., 1st sess., U.S. Government Printing Office, Washington, D.C., Vol. 1, 1963, 679 p., Appendix to Vol. 1, 1963, 508 p. [5S3] — House Committee on the Judiciary, Copyright Law Revision. Report of the Register of Copy- rights on the General Revision of the U.S. Copy- right Law, House committee print, 87th Cong., 1st sess., U.S. Government Printing Office, Wash- ington, D.C., 1961, 160 p. [5S4] — House Commitee on Post Office and Civil Serv- ice, Interim Report on the Use of Electronic Data Processing Equipment in the Federal Agencies, House rept. no. 627. 88th Cong., 1st sess., U.S. Government Printing Office, Washington, D.C., 1963, 4 p. [585] — Use of Electronic Data Processing Equipment. Part 1 — Comptroller General and Department of Defense, Hearings, 88th Cong., 1st sess., U.S. Government Printing Office, Washington, D.C. 1963, 194 p. [5S6] — House Committee on Science and Astronautics, Dissemination of Scientific Information. Hear- ings, 86th Cong., 1st sess.. U.S. Government Print- ing Office, Washington, D.C, 1959, 182 p. [587] — Dissemination of Scientific Information, Report of the Committee, House rept. no. 1179, 86th Cong., 1st sess., U.S. Government Printing Office, Washington, D.C, 1959, 12 p. [588] — Research on Mechanical Translation, Hear- ings, 86th Cong., 2nd sess., U.S. Government Printing Office, Washington, D.C, 1960, 183 p. [589] — House Select Committee on Government Re- search, Documentation and Dissemination of Re- search and Development Results, Study No. IV, 88th Cong., 2nd sess., U.S. Government Printing Office, Washington, D.C, 1964, 148 p. [590] — Senate Committee on Government Operations, Coordination of Information on Current Federal Research and Development Projects in the Field of Electronics, Committee print, 87th Cong., 1st sess., U.S. Government Printing Office, Washing- ton, D.C, 1961, 292 p. [591] — Coordination of Information on Current Scien- tific Research and Development Supported by the United States Government, Report no. 263, 87th Cong., 1st sess., U.S. Government Printing Office, Washington, D.C, 1961, 286 p. [592] — Documentation, Indexing, and Retrieval of Scientific Information. A Study of Federal and Non-Federal Science Information Processing and Retrieval Programs, Senate document no. 113, 86th Cong., 2nd sess., U.S. Government Printing Office, Washington, D.C, 1960, 283 p. [593] — Addendum to Senate document no. 113, Senate document no. 15, 87th Cong., 1st sess., U.S. Gov- ernment Printing Office, Washington, D.C, 1961, 22 p. [594] — Senate Committee on Government Operatons, Interagency Coordination of Information. Fed- eral Plans for Improvement and Use of Scientific, Engineering and Other Information Systems. Hearings, 87th Cong., 2d sess., Part 1 of 2, U.S. Government Printing Office, Washington, D.C, 1962, 242 p. [595] — Senate Committee on the Judiciary, U.S. Patent Office Research and Development Program, Corn- gram, Committee print, 88th Cong., 1st sess., U.S. Government Printing Office, Washington, D.C, 1963, 55 p. [596] U.S. Department of Commerce, Report of the Ad- visory Committee on Application of Machines to Patent Office Operations, (Vannevar Bush, chair- man), Washington, D.C, Dec. 22, 1954, 76 p. [597] — Office of Technical Services, U.S. Government Research Reports, Vol. 37, Cumulative Index, Semi-annual issue, Jan.-June 1962, 590 p ; July- Dec. 1962, 409 p., U.S. Government Printing Of- fice, Washington, D.C. [598] — ■ Non-Military and Older Military Research Re- ports, Vol. 37, Cumulative Index, Issues 1-24, Jan.-Dec. 1962, 150 p., U.S. Government Printing Office, Washington, D.C. (The Jan.-June 1962 cumulation is bound with its counterpart above and consists of 77 pages. ) [599] U.S. Department of the Interior, Bureau of Recla- mation, Thesaurus of Descriptors : a List of Key- words and Cross-References for Indexing and Retrieving the Literature of Water Resources Devolopment, tentative edition, Denver, Colo., Oct. 1963, 140 p. [600] U.S. Navy Department, Bureau of Ships, Thesaurus of Descriptive Terms and Code Book, Ruth D. Camp [comp], 1st ed., NAVSHIPS-250-210-1, Washington, D.C, Dec. 1963, 878 p. [601] U.S. Patent Office, Guide for Patent Draftsmen, U.S. Government Printing Office, Washington, D.C, May 1961, 16 p. [602] — 125th Anniversary of the United States Patent Act of 1836. Part II : Proceedings of the Inter- national Patent Office Workshop on Information Retrieval, U.S. Government Printing Office, Wash- ington, D.C, 1963, p. 77-231. [603] U.S. Public Health Service, Medical and Health Related Sciences Thesaurus, PHS publication no. 1031, Washington, D.C, Mar. 1963, 213 p. [604] Urquhart, D. J., A Review of the Results of the Royal Society Scientific Information Conference, 1948, ASLIB Proc. 4, 233-240 (1952). [605] Utman, R. E., Standards for Information Process- ing: a Progress Report, Computers and Automa- tion 12, 9-12 (1963). [606] Vann, J. O., Defense Documentation Center (DDC) for Scientific and Technical Information, J. Chem. Documentation 3, 220-222 (1963). 138 [607] Veilleux, M. P., Permuted Title Word Indexing: [630] Procedures for Man/Machine System, American University, Machine indexing : Progress and Problems, 1962, p. 77-111. [608] Veronese, U. H., UNESCO's Future Programme in [631] the Field of Libraries, Libr : 10, 1-9 (1960). [609] Vertanes, C. A., Automation Raps at the Door of the Library Catalog, Spec. Libraries 52, 237-242 [632] (1961). [610] Vickery, B. C, Classification and Indexing in Sci- [633] ence, Butterworth Scientific Publications, London, 1958, 185 p., 2nd ed. 235 p. [611] — Machines and Indexes, UNESCO Bull. Libraries 13, 249-252 (1959). [612] — On Retrieval System Theory, Buttervvorths, [634] London, 1961, 159 p. [613] — ■ Subject Analysis for Information Retrieval, in Proceedings of the International Conference on [635] Scientific Information, National Academy of Sci- ences-National Research Council, 1959, Vol. 2, p. 855-865. [614] — Thesaurus — a New Word in Documentation, J. Documentation 16, 181-189 ( 1960 ) . [636] [615] — The Tower of Babel: the Language Barrier in Science, in Library Association, Southport Con- ference, Sep. 19-23, 1955. Proceedings, Papers and Summaries of Discussions, 1955, p. 76-84. [637] [616] — The Universal Decimal Classification and Tech- nical Information Indexing, UNESCO Bull. Li- braries 15, 126-147 ( 1961 ) . [638] [617] Visscher, M. B.. The Interdependence of Knowledge and Information in the World Today, Lib. Quar- terly 24, 81-91 (1954). [618] Wagner, F. S., Jr., A Dictionary of Documentation [639] Terms, Amer. Documentation 11, 102-119 1960). [619] Wahlin, E., Principles for a Universal System of Classification Based on Certain Fundamental [640] Concepts and an Outline of a Variant Adapted to Technology, J. Documentation 19, 173-186 (1963). [620] Wall, E., An Action Plan for the Engineering Pro- fession, in American Association for the Advance- ment of Science, Information Systems — Essential [641] Tools in Engineering Application of Science for the Needs of Society, 1963, p. 16-20. [042] [621] — Action Plan for Improved Dissemination of Engineering Information, Progress rept. No. 1 (rev), Grant NSF-G-24836, Engineers Joint Council, New York, Nov. 23. 1962, 17 p. [622] — First Revision of the Thesaurus of ASTIA De- scriptors, Final rept. on Contract AF 18(600) [643] 1944, Engineers Joint Council, New York, Aug. 6, 1962, 19 p. [644] [623] — Information Retrieval Thesauri, Engineers Joint Council, New York, Nov. 1962, 17 p. [645] [624] — A Practical System for Documenting Building Research, in U.S. Congress. Senate, Committee on Government Operations, Documentation, Index- [646] ing, and Retrieval of Scientific Information, 1961, p. 180-203. [625] — A Unified Engineering Vocabulary for Use in [647] Information Dissemination Indexing, Storage, and Retrieval, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 1, 1963, p. [648] 37-38. [626] Walton, T. S., A Formal Indexing Language for Automated Document Retrieval Systems, in H. P. Luhn [ed], Automation and Scientific Communi- cation, Pt. 1, 1963, p. 21-22. [649] [627] Warheit, I. A., Bibliographic Identification and Or- ganization, Amer. Documentation 3, 105-110 (1952). [650] [628] — Evaluation of Library Techniques for the Con- trol of Research Materials, Amer. Documentation 7,267-275(1956). [651] [629] — Review of Centralization and Documentation, (Little, Arthur D., Inc report), Documentation Progress 6, 3-4 (1963). Waterman, A. T., The Future of Report Literature, in B. M. Fry and J. J. Kortendick [eds], The Production and Use of Technical Reports, 1955, p. 3-8. Weber, D. C, A Quagmire of Scientific Literature? College and Research Libraries 18, 103-106, 118 (1957). — The Road to a Bibliographic Utopia, D.C. Libraries, 27, 2-6 (1956). Weik, M. H. and V. J. Confer, Survey of Scientific and Technical Information Retrieval Schemes Within the Department of the Army, BRL rept. no. 1169, Ballistic Research Labs., Aberdeen Proving Ground, Md., July 1962, 95 p. Weil, B. H., Cooperative Information Processing — Petroleum, in J. H. Shera et al [eds], Docu- mentation in Action, 1956, p. 348-354. Weil, B. H., E. H. Brenner, K. Ornsen, R. Paddock, J. Raymond and S. Winn, Coordination of Docu- mentation Research and Information Services by the American Petroleum Institute, Amer. Docu- mentation 12, 56-60 ( 1961 ) . Weinberg, A. M., The Government Looks at Science Abstracting and Indexing Services, in B. A. Jen- sen [ed], Proceedings of Sixth Annual Meeting, 1963, p. 1-14. — ■ Scientific Communication, International Science and Technology, No. 16, Apr 1963, p. 65, 67-68, 71-72, 74. Weinstein, E. A. and J. B. Spry, Boeing SLIP : Computer Maintained Printed Book Catalogs, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 2, 1963, p. 233-234. Wellisch, H., New Publication Methods for the UDC, A Proposal, Rev. Int. Doc. 27, 145-148 (1960). Welt, I. D., A Combined Indexing-Abstracting Sys- tem, in Proceedings of the International Confer- ence on Scientific Information, National Academy of Sciences-National Research Council, 1959, Vol. 1, p. 449-459. — Indexes and Index Mechanization in Biomed- icine, J. Chem. Documentation 3, 169-174 (1963). Wessel, C. J. and W. M. Bejuki, Evaluation of Document Control in a Materials Deterioration Information Center, in National Academy of Sciences. National Research Council, Proceed- ings of the Internatonal Conference on Scientific Information, Vol 1, 1959. p. 731-761. Where the Computer Fits in Command and Con- trol, Armed Forces Management 8, 22-24 (1962). Whitby, T. J., National Bibliographv in the U.S.S.R., Lib. Quarterly 23, 16-22 (1963). Wichers, E., The Role of the Internatioinal Union of Pure and Applied Chemistry, J. Chem. Docu- mentation 3, 7-11 (1963). Wilkinson, W. A. and W. H. Waldo, New Informa- tion Services — A Practical Approach, J. Chem. Documentation 2, 175-177 (1962). Williams, E. E.. Farmington Plan Handbook, Assoc, of Research Libraries, Harvard Printing Office, Cambridge, Mass. 1953. Williams, T. M., Translating from Ordinary Dis- course into Formal Logic — a Preliminary Systems Study, rept. on Contract AF 19 (604)1582, ACF Industries Inc., Alexandria, Va., Sep.-Nov. 1956, 110 p. Wilson, C. D.. and G. A. Hamel, Developments in the European Translations Centre, Rev. Int. 30, 51-54 (1963). Winchell, C. M., Guide to Reference Books, 7th ed., American Library Association, Chicago, 1951, 645 p. Wofsey, M. M., An Analysis of Four COBOL Specifications, bulletin no. 5, Center for Tech- nology and Administration, American Univ., Washington, D.C, Mar 1962, 21 p. 139 [652] World List of Scientific Periodicals Published in the years 1900-1950, \V. A. Smith and F. L. Kent [eds], 3rd ed., Academic Press, New York; Butterworths, London, 1952, 1058 p. [653] Wright, M., The British Iron and Steel Industry Translation Service, ASLIB Proc. 12, 136-143 (1960). Conference on Technical Translation, London, 15 January 1960. [654] Wright, P., U. S. Computer Industry Can Capture European Market, Electronics Weekly, No. 156, Aug 28, 1963, p. 10. [655] Wright, W. E., A Report on Progress on Catalog Code Revision in the United States, Lib. Quar- terly 26, 331-336 (1956). [656] — Standards for Subject Heading: Problems and Opportunities, J. Cataloging and Classification 10,175-178 (1954). [657] Wyart, J., Scientific Documentation in France, in National Academy of Sciences-National Research Council, Proceedings of the International Confer- ence on Scientific Information, Vol. 1, 1959, p. 605-611. [658] Wyllys, R. E., Letter to the editor, Amer. Docu- mentation 14, 330 (1963) . [659] Yerkes, C. P., Microfiche, a New Information Media, in H. P. Luhn [ed], Automation and Scientific Communication, Pt. 2, 1963, p. 129. [660] Zusman, T. S., M. S. Thompson, J. B. Wilson, L. S. Rotolo and T. D. Gomery, Selected Bibliography of the International Geophysical Year: An Ex- ample of Tabledex Formats, National Biomedical Research Foundation, NBR and the Library of Congress, Rept. No. 62071/18100, Washington, D.C., July 1962, 109 p. 140 US GOVERNMrNT PRINTING 0FFICEI966