[Published in Alexandria, 5(1), 1993.]
By the 1970s it had become clear to most of those responsible for managing research libraries that neither funding nor available space could cope with the ever-increasing tide of publications, both in monograph and serial form, flooding from the world's presses. The recession brought about by the dramatic increase in the price of oil in 1973 added to the problem by increasing the costs of publishing while at the same time forcing governments to reduce expenditure on higher education. Solutions seemed few and reasonably obvious: encourage cooperative approaches by forming groups which could rationalize acquisition and preservation policies (the Research Libraries Group [RLG] in America and the Consortium of University Research Libraries [CURL] in the U.K. are two examples), and develop cooperative electronic cataloguing which would, hopefully, reduce the costs of maintaining manual catalogues and the wastage of manpower inherent in duplicate cataloguing effort. Many studies had proven beyond reasonable doubt that the maintenance of the catalogue in a medium-sized or large research library constituted a substantial proportion of the annual budget. In America and Canada we saw the rapid development of shared bibliographical resources like OCLC, RLIN, WLN and UTLAS, while in the U.K. there were initiatives like MERLIN (which collapsed in 1978), BLCMP, SCOLCAP, SWALCAP, LASER and UKLDS (which collapsed in 1982). There are plans, this year, to create a European consortium of libraries, modelled on RLG, to rationalise acquisition and cataloguing policies in the countries of the European Community. With the impetus given to the creation of national retrospective databases by the Eighteenth Century Short Title Catalogue [ESTC] soon to be combined with the catalogues for the period 1475 to 1700 (Pollard and Redgrave's STC and Donald Wing's STC, the revision of which is nearing completion), there are enough national projects currently in progress to encourage the view that a retrospective European database seems a practicable enterprise. With the improvements in international communication networks which make it technically feasible to access data held in remote databases it would seem that we are well on the way to achieving the IFLA objective of Universal Availability of Publications (UAP). Put another way, we seem to be on the threshold of successfully creating the "virtual library".
The consequences of this remarkable adventure, which would have appealed to the founding fathers of universal bibliography in the sixteenth century, are only beginning to be understood and it is the purpose of this essay to explore some of the problems which will have to be solved if knowledge really is to be enfranchised by the new technology.
It is a regrettable fact that in human history solutions to one problem can sometimes produce problems of another sort, usually orders of magnitude greater than anticipated: the consumption of the world's energy resources by developed nations to sustain economic growth has resulted in climatic changes which could well threaten the planet's survival. The enthusiasm with which libraries have adopted electronic devices to store information in the last ten years has not, I think, always produced the benefits which were promised and has certainly involved the expenditure of truly vast sums of money on hardware, software and telecommunications. The reasons for this are complex and this essay can achieve little more than to identify the nature of the problems and suggest a few guidelines for the future. It is probably fair to say that the overall objectives of an academic research library have changed little since Bacon's day: to support teaching and research[1]. The ways in which available resources are balanced generally depends on the emphasis in any given institution placed on one or the other and whether or not the institution is funded from private or public sources. With library budgets everywhere shrinking (in real terms) it is not difficult to see why the burden of economies tends to fall on expenditure supporting research: the monographs and serials are more expensive and they are less conspicuously used. In order to compensate for this, libraries increasingly look to remote access, both for identification of what is needed (online databases and their CD-ROM derivatives) and for document supply. Developments in optical disk technology now make it feasible to supply documents "on demand" electronically rather than in xerographic form, but it is noticeable that few libraries are prepared to provide these facilities as freely as they have traditionally provided printed materials. No library I know of charges students or staff for access to the printed National Union Catalog (formerly maintained as a card file at the Library of Congress until its destruction in the 1970s), even though its production required an enormous investment by the Library of Congress and by the libraries throughout the world that agreed to subscribe to it. It is a prodigious - if flawed - research tool, the flaws deriving in large measure from the methods used to translate a vast card file of duplicate entries, with widely varying sophistication of bibliographical description, into a printed catalogue based on the principle of unit entry with locations attached. Because this task was entrusted, for the most part, to unskilled staff an enormous amount of valuable information about individual copies of early printed books was allowed to disappear without trace.
The publication of NUC was greeted with enthusiasm by libraries everywhere: here at last was a palpable contribution to the dream of a universal bibliography first articulated by the Society of Arts in the 1850s. The costs involved in acquiring and providing access to this vast inventory were seldom questioned, even if some reference librarians complained about the necessity to relegate other bibliographical tools to less conveniently situated locations in their library. At about the same time, the firm of G.K. Hall embarked on an ambitious programme to publish in volume form the card catalogues of specialist libraries, thereby putting even greater pressure on libraries to find room for open-access accommodation for research tools of primary importance. That pressure had become so considerable by 1980 that it was no longer possible to shelve such basic reference works in the Reading Room of the British Library and a large number of them were consigned to the stacks - hardly the place to put library catalogues!
I have heard it said that the development of large online cooperative catalogues has made the multiplicity of printed library catalogues largely redundant. That, as anyone who has used these catalogues will testify, is far from true. A library's responsibility for describing the books and manuscripts in its care is quite different from that associated with a database serving a widely dispersed community of collaborating institutions. For obvious reasons large databases discourage contributing libraries from including what is regarded as "local" (including copy-specific) information. Thus, while one can discover from OCLC and the printed catalogue of the Research Division of the New York Public Library the existence of certain items by T.S. Eliot, it is nevertheless necessary to consult the catalogue of the Berg Collection (published by G.K. Hall) to discover their real importance. Even ESTC, from the beginning conceived of as a scholarly resource and therefore willing to include "local" information where it has been supplied, cannot hope to do more then point scholars in the right direction. ISTC, while seeking to list surviving copies of fifteenth-century printing in every country nevertheless provides scant information about specific copies, assuming (rightly I believe) that if it is necessary to know more about an identified copy of a book printed by Ulrich Gering in the Bibliothèque Nationale then the place to find that information is in the printed catalogue of that library's collection of incunabula. Thus, ISTC is silent on the existence of Melanchthon's copy of Ovid with notes and drawings by him in the Brotherton Collection at Leeds University, though the existence of a copy there is noted. It may be a source of profound melancholy for those who believe that we are entering a "new age" in which electronic information will gradually supersede print but it remains a fact that if research has a future then both electronic and eye-readable resources must have their place in the research library for the foreseeable future.
If this proves to be true then it should follow that one kind of resource should not be discriminated against in favour of the other. As far as the researcher is concerned it does not matter how knowledge is communicated, but it does matter whether or not there is an entitlement to it. For the first time in history librarians, desperately seeking to discover some mechanism to increase revenue, have put forward a case for charging users with a subtle form of information tax. The illogicality of granting a reader free access to NUC, and quite literally kilometres of similar printed resources, but charging for access to electronic databases is extraordinary. The cost of acquiring and maintaining a good representative collection of printed and microform catalogues almost certainly exceeds the investment required for electronic access to remote databases, especially if such access is undertaken by a skilled interpreter. Creating a generation of young librarians with the skills needed to perform such a vital function seems to me one of the most important missions of library schools, but for such a mission to succeed it is obvious that there must be an understanding of four fundamental elements in the "information process": (1) the nature of the materials described (whether printed, manuscript or audiovisual); (2) the variety of cataloguing conventions adopted to describe such materials; (3) the adaptation of these conventions to database tagging; (4) the variety of syntactic protocols for interrogating tagged files. The first two elements belong to the traditional discipline we used to call "historical bibliography", the others to "information technology". That both a historical and technical understanding is required to use the resources available to today's researchers comes as no surprise to my academic colleagues in disciplines like English and History but it seems to have escaped the attention of those planning the curricula of library schools who would appear to value IT skills above all others. The library in which a student or researcher is provided with free and unimpeded access to bibliographical information held in remote databases like OCLC, RLIN or BLAISE simply does not exist. Nor, I fear, is it likely to exist in the foreseeable future: and for a number of reasons. One very obvious reason has to do with the retrieval devices needed to access electronic information. While it is true that catalogues take up a great deal of space, they have the advantage of permitting simultaneous consultation by as many users (theoretically) as there are component units (drawers of cards or volumes). At the busiest time of the day in the Reading Room of the British Library (3 p.m.) there can be as many as fifty readers consulting the 2008-volume catalogue. Given that most of these readers are in search of books by known authors or listed under known headings the period of consultation is generally short - on average no more than two minutes. A recent survey has shown that in an average day the Reading Room catalogue is consulted about two thousand times. By contrast, readers using the six OPAC terminals provided for access to books published or acquired since 1975 spend an average of fifteen minutes per search. This is due to one of two reasons: readers have difficulties in understanding how the system works; the provision of enhanced subject and keyword access results in searches producing a large number of records which must be browsed. Observation suggests that the multiplier which must be used when access to bibliographical information is transformed from manual to electronic is a factor of seven. There are no reliable studies of this phenomenon for the simple reason that only a handful of libraries anywhere in the world have converted the catalogues of their total stock to machine-readable form and made provision for free access. Most of the converted catalogues which can be seen in universities in North America and Europe are for modern books and serve, primarily, undergraduate needs. As James Thompson has observed, serving undergraduate needs is neither complicated nor vastly expensive:
Even if we extend this figure to about 250,000 volumes, providing straightforward author/title/subject access does not today present daunting problems as there are several hardware and software packages available which will perform the task efficiently and cost-effectively. Such a stock, distributed evenly over the range of subjects taught, will be found to yield about ten `hits' for an author search, fifteen for a subject search and twenty for a title keyword search. Deciding on which titles seem appropriate to the student's needs will normally require no more than five minutes at the terminal because most terminal display systems can accommodate between twelve and twenty brief titles per screen. Increasing the size of the database does not, however, increase the time needed arithmetically: once a database exceeds a million records the time needed is closer to a logarithmic progression with the time needed increasing threefold for every doubling of size.
Observation of readers using the OPAC in the British Library reveals that individual searches, frequently leading to no more than three applications for books, currently occupies about fifteen minutes. If, as is planned in the new British Library building on Euston Road, readers are given OPAC access to the full range of databases available on BLAISE-LINE, experience suggests that there will have to be at least 200 terminals provided in order to satisfy initial demand by staff and readers. As readers become familiar with the system, and the wonderfully complex searches made possible, demand will undoubtedly increase. More important, perhaps, is the fact that complex searching makes heavy demands on processing power, as all users on online catalogues know. With no more than fifteen users simultaneously logged on BLAISE-LINE a search involving a single truncation and one Boolean operator (and/or/and not) can force the user into making ten keyboard responses to the terse request "CONTINUE PROCESSING Y/N"! Experience in America suggests that library administrators invariably underestimate demand when a library catalogue is made available electronically. Apart from the Library of Congress REMARC file (based on the shelf-list) - and for which access is charged on DIALOG - no library automation project can match, for size and complexity, the conversion of the British Library's General Catalogue recently completed. It is accessible in two forms: the online version, available to registered users of the BLAISE-LINE service, and the CD-ROM version on five disks, marketed by Chadwyck-Healey Ltd. Users in the Round Reading Room have access to the online version in the Research Seminars I give on Tuesday and Thursday mornings, and the CD-ROM version is available on one microcomputer. The two versions are complimentary, as some searches are possible only on the CD-ROM version. The CD-ROM version of the General Catalogue to 1975 (with over six million records) was made available to readers for the first time on June 20 1992 and observations suggest that for most searches readers retrieve more records than they can handle. Because of the way in which the software functions, looking at full displays for more than twenty records becomes a tedious operation. If the search involves a complex heading like England or London (or a major author like Shakespeare or Goethe) the time needed to find a record can become unacceptable. Because of the way in which the General Catalogue was keyed - and subsequently indexed - the facility to browse a heading is impossible. You may find the record you searched for somewhere in the middle of ENGLAND - Appendix - Topography and Travel - but you have no way of seeing what other titles are in the vicinity of that record. This, as many a user of the General Catalogue has testified, is one of the most useful and instructive features of a systematic catalogue whether compiled on the dictionary principle (the printed catalogue of the Research Collections of the New York Public Library is a good example) or the encyclopaedic principle (like the General Catalogue). Researchers are well aware of the fact that the sources on which their research depends have associations which are extraneous to their bibliographic descriptions and the ability to browse in one of the many chronological arrangements within the vast ENGLAND heading in printed versions of the General Catalogue are simply not matched in either the online or CD-ROM versions. Given the procedures associated with adjacency (books on the shelf, entries in a catalogue) in discovering unknown sources this seems to me a retrograde development. There is no software currently available which permits adjacency browsing (previous record, next record) on an online catalogue, because very few catalogues have been converted sequentially. Yet the British Library's converted catalogue was converted sequentially and each record was provided with an identifying number which reflects its position in the column, the page and the volume. Except for the supplementary volumes which include all the records which were omitted (for a variety of reasons) from the initial conversion it would have been a fairly straightforward matter to invert the online catalogue on the record number, thereby making it possible for users to be presented with the first record in a heading rather than the last. The additional facility of providing for "up record" [the minus key] or "down record" [the plus key] would then have made adjacency searching feasible. One example will suffice to illustrate my point. Users of the General Catalogue in search of editions of Justinian's Institutes are notified that these are to be found under the heading ROME [Emperors.] - JUSTINIAN I., etc. [527- 565.] This is a very complex heading with subheadings for CORPUS JURIS CIVILIS (with sequences for the text, Index and Appendix); CODEX AND DIGESTA; CODEX AND Vetus NOVELLÆ CONSTUTIONES; CODEX; DIGESTA; (with sequences for Digestum, Infortiatum, Digestum Novum, Extracts and Selections, Single Titles, Index and Appendix); EDICTA; INSTITUTIONES (with separate sequences arranged by language); NOVELLÆ CONSTITUTIONES (with separate sequences for Epitome, Selections and Appendix); MISCELLANEOUS; and APPENDIX. Occupying just 56 columns of print, this heading is regarded as "minor", yet it is impossible - using either version of the converted catalogue - to get any sense of the immense intellectual effort which underlies its structure, just as it is virtually impossible to construct a search which will retrieve only the English texts of the Institutes rather than the English texts of the Digesta. If getting some sense of the bibliographical history of Justinian is difficult in the converted catalogue it is frankly impossible with major authors like Luther. One of the principal reasons why electronic catalogues can be difficult to use is that they manifest all the signs of what happens in a developing technology. When MARC was developed as a tagging system for bibliographic files in the 1960s it was assumed that libraries would want to have their traditional card catalogues maintained for the foreseeable future, and so elaborate rules were devised to enable utilities like OCLC to format card output for electronic records. In the British Library the development of software to produce the British National Bibliography was undertaken and a more sophisticated version of that software was used to produce the microfiche versions of the Library's post-1975 holdings as well as ESTC in 1983 and 1992. One irritating feature of most online catalogues is the fact that no distinction in searching is possible between a name as a main entry (100xx) and the same name as an added entry (700xx) any more than uniform titles (240xx) can be distinguished from titles proper (245xx). Thus, a search for works by Martin Luther on either OCLC or the converted BLC file will also retrieve the records for works with a contribution by him (e.g. on OCLC) or the cross-references under the Luther heading in the General Catalogue. A welcome feature incorporated by Saztech in the CD-ROM version of the converted BLC is the subfield tagging of the imprint field which makes it possible to perform chronological searches, not possible on the British Library online version. But readers are puzzled by the fact that a search for all titles in a heading for a certain year retrieves records with quite different years in the initial display of brief records. This is quite easily explained, of course, and readers are getting used to the fact that the converted catalogue has both "unit" and "blocked" records, and most users of the Library come to the electronic versions of the catalogue with some experience of its printed version. But what will happen when the Reading Room Catalogue has to be withdrawn from use and readers will be expected to find what they need only with electronic versions? And what about remote users who may, in the near future, have access to these electronic versions and who have no understanding of the way in which the catalogue has developed? These are questions of some importance which no-one, it seems to me, is addressing.
It can be argued, of course, that catalogues like those available in Bloomsbury or New York are eccentric creations and that most library catalogues follow the rigid prescriptions for author/title entries. It is probably just as well, in America at least, that this is so, otherwise it would be difficult to see on what basis a union catalogue could be constructed. The widespread adoption by American and Canadian libraries of AACR2 has, if it has done nothing else, made it possible to create vast databases like OCLC and RLIN in which records from a wide variety of contributing libraries (with widely differing standards of cataloguing practice) can be assimilated. However, there are two important factors which make these databases difficult for inexperienced users. The first has to do with the competence of the libraries contributing data; the second with the competence of the users. If records entered on a union catalogue were to conform to internationally agreed standards in (1) the adoption of headings (principal and subsidiary); (2) physical description; and (3) subject indexing, many of the problems which users encounter would disappear. Unfortunately, there are still huge unresolved areas of disagreement between even the national libraries in English-speaking countries, leaving aside for the moment disagreements between them and the rest of the world. Reconciling retrospectively the differences, for example, between Britain and America on the form of name appropriate for nobility is unlikely ever to be undertaken, and users must appreciate that fact. It is generally thought that retrieving an author from a bibliographic database is a trivial matter. That this is not so can be simply demonstrated by trying to find Henry Smith (Professor of Languages in Marietta College, Ohio) the author of a Greek and English lexicon to the works of Homer, first published in 1844, on the online version of BLC (the General Catalogue of the British Library to 1975): the author index reveals 482 entries for SMITH H and a search on author words (aw) for "Smith" and "Henry" yields 1039 entries. It is no easier to find this author on OCLC! If you know the exact title of a work then it is frequently more efficient to search on keyword, but in this case a search on BLC for greek [and] english [and] lexicon retrieves 55 entries, while the same search on OCLC retrieves 321 entries. If, as is often the case on large collaborative databases, there are differing forms of the name (including different dates of birth and/or death) then the user has no option but to browse the author index (a crucially important facility provided by most systems). Although OCLC is making considerable efforts to try and tidy their database and remove obvious duplicates there still remain a huge number of variant entries for well-known authors. For example, the subject index for T.S. Eliot has the following [note that spaces are crucial]:
The personal name phrase index has, on the other hand:
On neighbouring screens we find:
I find it hard to believe that Aristophanes and Aristotle can cause so many difficulties to cataloguers:
The difficulties created by the lack of uniformity in establishing headings are not trivial, since it necessitates retrieving records that might be relevant to one's search - a procedure which takes time and costs money. Users of online databases are just beginning to realise that subject indexing is not quite the aid to research that it has been made out to be. This has less to do with the principles than their application. Apart from the quite obvious advantages of a numerical system in an electronic environment (performing truncated searches on Dewey numbers can often smooth out differences in finely tuned hierarchies) the use of Library of Congress Subject Headings (LCSH) is pervasive on both sides of the Atlantic. But for a lexical system to work effectively its rules must be followed scrupulously, and this is virtually impossible to achieve in a collaborative database that lacks editorial control (such as is provided for in ESTC or ISTC both of which projects maintain strict editorial control over what is entered). If users experience difficulties with the rules governing subject headings large databases like OCLC provide some novel problems. Starting as a catalogue utility serving the state of Ohio, OCLC has now become an international enterprise with an increasing number of participating libraries outside the United States and Canada. Not too many users are aware of the fact that libraries in Europe are putting their records on the system using non-English subject headings for books in languages other than English. The volume of such records is not as yet substantial, but it is likely to increase. And if the plans for a European database, modelled on RLIN (the database maintained by the Research Libraries Group in America), succeed then it will presumably be expected to accommodate vernacular conventions of cataloguing. Those attending my research seminars are frequently surprised by the fact that some French books on French history are not retrievable using English subject terms, and some books on Padua can only be found by searching on Padova. Differences between national standards for the form of personal names can lead to separate forms in OCLC for Anthony/Sir Anthony/Antonio Panizzi. For Gottlieb Christian Crusius, author of a Greek and English lexicon to the works of Homer, the following forms are found:
But few names seem to have the variety of forms found for Virgil. Many of the University of London catalogues (which use Libertas) display variations in the name, but no system can match OCLC which has the following:
For Peter Carter's Offences of Violence (London, 1991) the OCLC subject terms are: Infractions contre la personne - Angelterre; Violence (droit) - Angleterre. Jean-François Gournay's L'Érotisme en Angleterre: XVIIe-XVIIIe siècles is to be found in only one record with the subject terms: Erotisme - Angleterre - Histoire. Dorothy Van Woerkom's Perle et les Menestrels (Paris, 1983) was entered by a French library with the subject terms: Menestrels - Romans; Angleterre - Romans.
For the past three years I have been providing assistance to
readers at the British Library with their research projects. These research
seminars have now become a familiar extension to the service traditionally
provided by the staff of the Enquiry Desk, and though the number of readers
taking advantage of the service is relatively small, it is consistent. In an
average week I receive five requests for help, and the list of projects covered
(in all historical disciples) now exceeds six hundred. The seminars are not
seen as a primary resource but rather as an ancillary aid for those
encountering real difficulties with primary and secondary sources and who need
access to bibliographical information only available electronically, the volume
of which increases annually. Significant features of a file like OCLC are: its
currency; the number of dissertations and microforms catalogued; and the access
it provides to manuscript and audiovisual and other non-book sources. On the
other hand, specialised files like ESTC and ISTC,
editorially controlled, well indexed, and covering the holdings of a very large
number of libraries, provides an unrivalled source for sophisticated searches
of a kind undreamt of a decade ago. While ISTC is structured and indexed in such a way
as to provide those concerned with fifteenth century books with every possible
finding aid - in fact the database is a remarkable example of a bibliographical
finding aid rather than an online bibliographical catalogue based on AACR2 - one must
understand the cataloguing rules on which it is based in order to achieve the
best results from a search strategy. This is particularly important for a
database such as ESTC,
since it is a sometimes uneasy compromise between bibliographical principles
and the prescriptions of AACR2.
The diversity of cataloguing codes evident in virtually every printed catalogue
or bibliography from Trithemius[2]
to the present day does not usually present great difficulties for experienced
researchers: a cursory inspection of the introduction, the arrangement and the
indexes (if any) are generally enough to provide the clues necessary to
determine its morphology. The General Catalogue of the British
Museum Library is an exception to this general rule, for only once in its
remarkable history since 1787 has it been published with any guidance to users
as to how it was constructed: the aborted edition of 1841, of which only one
volume appeared, with Panizzi's celebrated "XCI Rules" prefixed. Its final
manifestation in published form (generally referred to as GK4 or BLC)
and in the laid-down set in the Reading Room provides no guidance to users on
the changes which have taken place over the years with respect to establishing
headings and providing cross- references. Every generation of new users must
learn anew the rules for entering periodical publications or journals of
learned societies, anonymously or pseudonymously published books, books in
series, and the rules governing entry for complex headings like BIBLE, ENGLAND, JESUS
CHRIST, LONDON, ROME and UNITED STATES. It may take years for even
persistent users of the catalogue to discover the riches in headings such as CATALOGUES or COLLECTION.
The variety of morphologies found in printed catalogues is no less conspicuous in electronic databases, and there are indications that if the future of research libraries is destined to become ever more inextricably involved with electronic information one of two courses of action must follow as a consequence: either the users must be educated, or libraries will have to provide a new breed of reference librarian equipped with a detailed knowledge of data structures, cataloguing rules, telecommunication and search protocols as well as convenient public access to the networks providing access to data remotely held: BLAISE-LINE, INTERNET, BITNET, JANET/SUPERJANET[1], &c. Which of these solutions is likeliest to be adopted is a question which cannot be answered and is probably one of the most difficult problems (apart from shrinking resources) which librarians must address in the very near future. That the education, even re-education of users was a necessary consequence when computers were introduced into libraries in the 1970s was obvious enough, and since 1970 the literature has expanded sufficiently large to justify a bibliography: Judith M. Pask's User Education for Online Systems in Libraries: a Selective Bibliography 1970-1988[3]. Although this bibliography is understandably highly selective, omitting works like Anne Beaubien's Learning the Library[4] and Helen Wheeler's The Bibliographic Instruction-Course Handbook[5], it nevertheless lists 555 books and articles. In the Introduction she writes:
One of the major problems for librarians designing user education is that online catalogues vary so much between libraries - even more than card catalogues did. Computer technology has given librarians a wide variety of options in adapting their catalogues to unique local needs. In fact, many systems began or evolved as a tool for librarians to manage or organize the collection, not primarily as a library users' tool. Several librarians have written about the need for bibliographic instruction in general and online systems training in particular, to change from a procedural (step by step) approach to a more conceptual approach. through better understanding of the organization and control of online systems, users are able to apply these skills to their individual needs, to more complex searches, or to different systems. This also corresponds with current trends in education. Emphasis now is on "process" rather than "product" - on the pupil acquiring a capability for learning how to learn, rather than on just becoming learned. this new approach will not only increase library use, but will also create a climate for long-term library instruction - that is, teaching the process for finding information, rather than just teaching how to use a specific tool or specific online system. If my own experience with online catalogues (starting with the creation of ESTC in 1976) is relevant then it seems to me that anyone whose research depends upon access to a multiplicity of databases will need to understand how electronic catalogues are constructed in order to cope with the problems of what exists and where it can be found in the years ahead, unless research libraries are prepared to provide experienced staff to assist them in their research. Since the demise of the scholar librarian about thirty years ago researchers have shown little interest in the philosophical, sometimes theological, debates which enliven meetings between librarians. The discipline of bibliography - what De Bure[6] called "the description of books" - has now all but disappeared to be replaced by informatics, much as criticism has yielded to hermeneutics. But the cataloguer, in whatever medium, must address the same problems today as yesterday, and will continue to do so as long as we have objects called books, with features inherited from a tradition extending back a thousand years.
But what of information which is itself electronic and exists in no other form? A rapidly growing corpus of electronic "records" (deriving from government, commerce, institutions and private individuals) as well as digitised image-files of perishing printed and manuscript sources destined to disappear because we do not have the resources to preserve them for posterity will soon become legitimate targets for historical investigation. They too must be listed and described, but what format will we use to describe them? I have no doubt that some future version of AACR will take this into account and that MARC will be pressed into service to provide the necessary codes (as it has done for microform and other forms of non-book publishing), and researchers will have to master a subsidiary set of rules[2]. If some of the views expressed in Constance Mellon's anthology (Bibliographic Instruction: the Second Generation)[7], prove correct, then I think it is arguable which is the most cost-effective route for libraries to take: to educate users for every generation to come, or provide a small cadre of interpreters (combining informatic and hermeneutic skills) to guide users in their search for bibliographic information. Either way, it seems to me that research libraries - if they are to survive - are bound to re-establish that collegial relationship which has always sought to make libraries institutions which bring together the enquiring mind and the materials on which enquiry thrives.
Faced with the bewildering variety of file formats found in online catalogues it is hardly surprising that, from time to time, one hears of schemes to resolve the problem by either persuading database providers to adopt uniform practices (the solution suggested by William Buckel[8]) or to develop heuristic software which would enable any microcomputer to interrogate any database using a standard query language. Either of these ingenious and easily understood solutions is about as likely a development as a common standard for headings shared by the national libraries of the English-speaking peoples!
In the long and slow process which has brought us from darkness to light the role of the library has always been crucial: both as custodian of knowledge and information and a provider of the means whereby they can be accessed and used to advantage. The history of "bibliography" (interpreted as de Bure did in the widest sense) cannot be appreciated without reference to the history of libraries, and the dissociation between libraries and those who use them, if it is being brought about by the appearance of a new kind of research tool, must be seen as a retrograde development and one which threatens a vital collaboration which has got us where we are. In time, when schoolchildren learn how to handle bibliographical databases and online catalogues have help files which prompt users with intelligible guidance on searches and try to interpret their mistakes rather than display bizarre error messages, the need for interpreters may diminish. But there will always be, I believe, a conspicuous need for librarians who understand the nature of the collections in their care (what they contain and what they lack); the peculiar manner in which these collections are described; and an understanding of the devices at their disposal for providing users with access to them.
1. For a wide ranging series of essays on academic libraries and their future see: Maurice B. Line (ed.), Academic Library Management. London, Library Association, 1990.
2. Ref. 1, p. 15.
3. Hancock-Beaulieu, Micheline & Stephen Walker: “An evaluation of automatic query expansion in an online library catalogue”. Journal of Documentation, 48 (4), 1992, 406-21. Note also: C.L. Borgman, “Why are online catalogs hard to use?” Journal of the American Society for Information Science, 37 (3), 1986, 387-400.
4. Walker, Stephen & Micheline Hancock-Beaulieu. Okapi at City: an evaluation facility for interactive IR. London, British Library, 1991 (R&D Report 6056).
5. Yates, Frances A. The Rosicrucian Enlightenment. London, 1972. BL subject heading: Rosicrucians.
6. Yates, Frances A. Giordano Bruno and the hermetic tradition. London, 1978. BL subject heading: Occult sciences – History; OCLC – Occultism – History.
7. Ginzburg, Carlo. Il Nicodemismo. Torino, 1970. OCLC – Nicodemites.
8. Jacob, Margaret C. The Newtonians and the English Revolution, 1689-1720. Ithaca, 1976. BL – Religion and science – England – History; Latitudinarianism (Church of England) – History.
9. Yates, Frances A. The Art of Memory. Harmondsworth, 1978.
10. Partridge, Chris. “Got a question? Just ask a surfer.” The Times, 6 November 1992, p. 10.
11. Quoted in ref. 10.
12. Often referred to as “the father of bibliography.”
13. JANET currently provides access to over 50 university and academic library catalogues.
14. Pask, Judith M. User education for online systems in libraries: a selective bibliography 1970-1988. Metuchen, 1990.
15. Beaubien, Anne. Learning the Library. London, 1982.
16. Wheeler, Helen. The Bibliographic Instruction-Course Handbook. Metuchen, 1988.
17. De Bure, Guillaume F. Bibliographie instructive. 10 vols. Paris, 1763-82.
18. Archivist have led the way here and dealing with electronic records is now an established part of archival management. Chalres M. Dollar, Electronic Records Managament. Paris, 1986.
19. Mellon, Constance. Bibliographic Instruction. Littleton, 1987.
20. Buckel, William. “The Uniform Catalog”. Library Journal, 111 (1), 1986, 52-54.
[1] JANET currently provides access to over fifty university and academic library catalogues. In almost all cases users are prompted with simple and easily-understood menus. It is not always easy to browse the indexes, but many systems will display index terms adjacent to the desired word if the request is unsuccessful. Registered users of INTERNET have access to roughly the same number of university library catalogues in America.
[2] Archivists have led the way here and dealing with electronic records is now an established part of archival management. See further: Charles M. Dollar, Electronic Records Management and Archives in International Organizations: a RAMP Study. Parus, UNESCO, 1986; Margaret L. Hedstrom, Archives & Manuscripts: Machine-Readable Records. Chicago, Society of American Archivists, 1984.