Cataloguing and indexing local studies material: looking forward from 1983

 Cataloguing and indexing local studies material:

a paper presented to the Library Association Local Studies Group post-conference seminar

“Access to local studies materials”

held at Torquay 22nd to 23rd September 1983

by Ian Maxted

Introduction

Eleven leap years on, as I am at long last finalising work on a local studies thesaurus for the Devon bibliography, I chanced upon the typescript of this paper. At the time, several people expressed regret that it had never been published; today they may express regret that it had never been thrown away. Reading it through one last time before doing so, I thought it shed an interesting light on local studies librarians coming to terms with what was then “new technology” in the early 1980s, hence the frequent mention of card indexes. So much has happened since then in the field of search engines; AACR2 has been replaced by RDA as a metadata standard and artificial intelligence is on the horizon, but has so far been of no help at all to me in my bibliographical endeavours. I present the talk here with very few changes, expanding some of the abbreviations that may not have stood the test of time, updating the tables to some extent, and cutting down on my excursion down the blind alley of pronounceable notation. And the ideas being arrived at by me in 1983 did form the basis for subject indexing in the Devon bibliography over the past four decades.

Leap Year’s Day 2024

--oOo--

Local studies libraries are different. Normal everyday librarians look on in disbelief as we rummage through displays of leaflets to add to our collections of ephemera, or smile tolerantly as we write off for microfilms of obscure documents in the Public Record Office or Lambeth Palace Library. After all, they say to themselves, it represents only a small proportion of the total book fund and it will presumably be useful to researchers in the future. They somehow never imagine that it could also be useful to researchers in the present. And so we are left to go own way and to organise our collections in a manner that is frequently different from that adopted by the rest of the library service.

And it must be admitted that the contents of local studies collections do differ in many ways from the general public library’s stock. Multimedia resources are nothing new. As early as 1800 S. W. Fores, the London print-seller, was lending out portfolios of engravings for the evening, and by 1859 Gray and Warren of Croydon were running a stereoscopic circulating library. Though the idea may have taken some time to catch on, non-book materials are taken for granted in public libraries today, but they are still not integrated as they have been for many years in the local studies collection.

The local Studies librarian will, without a second thought, range across books, periodical articles, cuttings files, illustrations, ephemera, maps and manuscripts to answer a question on the history of a particular house or local worthy. Even within the category loosely described as books there is much that is not caught by the net of national bibliographies or other databases. Report literature, newsletters, publicity brochures, timetables, sale catalogues, prospectuses, by-laws, minutes, agendas are all grist to our mill, most are retained permanently, many are catalogued or indexed in detail.

Equally different are the means by which collections are built up. Donations figure larger than in other library departments; indeed local collections can be named after or based on the collections of a local antiquary or collector who has left his library to the town or county. The order in which this individual arranged and listed his collection may be a significant element in its research or archival value, and this is often recognized and retained in the local studies library. Local organizations may deposit material, for example the local photographic society may conduct a photographic survey or the local radio station deposit tapes of programs. Such parts of the collection may demand special treatment so that they can be compatible with the needs of the depositor, who may require immediate access to them. Some donations are actually conditional on the library keeping them separate. Add to such circumstances as these, periodic reorganizations of local government which can merge or split collections, and the web can become very tangled over time.

So, just as each locality is different, each local studies collection is different too, and the means of organizing the stock evolved by the custodians of each collection are equally varied. Arrangement of the stock ranges from virtually complete open access to virtually complete closed access, from fixed location to unaltered Dewey. Stock records range from complete and painstaking cataloguing to perfunctory stock lists. The form of the catalogues too can vary from printed catalogues with an infuriating series of supplements to the traditional card catalogue, and the main catalogue may or may not contain periodical articles, guidebooks, illustrations and other categories of material. Subject approach may be classified or alphabetical, place or subject may have priority, or it may even be necessary to rely on the grey head of a long-serving member of staff to obtain detailed subject access. And even if the books have been treated in a relatively orthodox manner, one cannot always be prepared for the very individualistic approaches to the problems posed by illustrations, cuttings or maps.

A half-hour talk is no place to enter further into the wildering range of possibilities. Suffice it to say that many researchers who are hardy enough to attempt comparative local studies find the lack of consistency from one library to another a considerable problem. At times the different approaches to organising collections mean that direct comparisons between localities are rendered difficult if not impossible. It is perhaps this lack of standardization which may lead those in the library authority who are considering computerization to leave the local studies collection to one side. The reaction from the local studies librarian may be a sigh of relief and a smug reflection that, if computers cannot cope, the job of organizing a collection such as this must be a truly professional one and the founts of all wisdom who occupy the dusty niches in the local history room will remain indispensable for a good few years yet.

But computerization will come. In some local collections it already has come, and some authorities even before the advent of the microcomputer had begun to computerize the catalogues of sections of the local studies stock. But will the future simply bring us a computerized version of the present fragmented situation? The early years of any new development have to be a period of experimentation. But in ten years time will the end-product of all this research simply be the ability to achieve in nanoseconds the anarchy which it has taken legions of devoted local studies librarians decades to arrive at?

Local studies librarians must be able to face the new technology in a united manner. True, each collection and each community differs, but surely there are enough common features to form the basis of a reasonably consistent application of computer technology. Let us try to be in a position to say what we want from computerisation. The rest of my talk will try to identify some of the factors involved in arriving at this position and I should stress that, like so many local studies librarians, I am talking as someone who has no detailed knowledge of computer applications or of modern indexing techniques. I have no hands on experience but perhaps it is no bad thing that someone who cannot tell a bit from a byte can present a lay person's view of what he hopes for from computers. We are all here to learn, myself as much as anybody, and if somebody can stand up and say that such and such a package is tailor-made for our needs, I am sure that we shall all be thankful, if a little sceptical.

Let us consider some of the problems involved, some of which have been alluded to already.

Firstly, the problem of the diversity of materials. Ideally we should be looking for some type of system that is equally capable of handling the multi-volume county history and the individual postcard or broadsheet, if necessary in one sequence but in practice normally in separate subsets.

Secondly, the problem of arrangement. The value of local studies material has risen considerably over the years and even the libraries which practice the most liberal policy of open access have their own securicorner. Also the wear and tear on less valuable items through continuous browsing has led to their withdrawal from the open shelves or replacement by photocopies or microfilms. The archival nature of many collections has led to their being shelved together regardless of subject content. Early pamphlet or tract collections are often physically inseparable because they have been bound together in volumes. For one reason or another therefore, classified shelf arrangement does not play the key role that it does in the open access lending or reference library. Subject access has to be increasingly through catalogues and indexes. It is more important to be able to pinpoint the individual item required than to put one's hand on a selection of broadly similar volumes on the same shelf.

Thirdly, there is a problem of the smallness of most local studies departments. Even where there is a separate staff they are a few in number and qualified and unqualified members have to play their part in the whole range of duties, both professional and non-professional. This is as true as of indexing as of any other activity. Thus any system should be simple enough to operate with a minimum of training and fields, tags, or other operating devices being kept as straightforward as possible.

Fourthly, the problem of technological lag. As already mentioned, there will be a delay in introducing computerized systems into many local studies collections. It is important therefore that the basics of the system should be amenable to manual application as a prior step to computerization.

Bearing such constraints as these in mind what therefore should one look for in a standardized retrieval system for local studies materials?

Firstly, the format of the entries. Over many years much international discussion has evolved AACR (Angle-American Cataloguing Rules, 1967) as a standard for bibliographical cataloguing. It is now in its second version (AACR2, 1978), and will doubtless be modified further but it is not likely to change in its essentials. It is capable of dealing with a wide range of materials beside monographs, and entries can be as full or as brief as one wishes to make them. It is also the basis for entries in the major bibliographical databases. AACR2 [now replaced by RDA Resource Discovery and Access, 2010] therefore should form the basis of any catalogue in local studies collections.

With regard to the choice of heading for corporate bodies, I would make some changes of emphasis, largely to cope with the semi-archival role of local studies collections.

Firstly I would prefer the straightforward form of the name of the corporate body in natural language: Devon County Council, for example, rather than the contorted forms produced by AACR Devon (County). Council.

Secondly subsidiary bodies should be more clearly shown as an integral part of the main body. Thus where the British National Bibliography might use the form Devon County Tourism Office, for local studies collections the form Devon County Council. Tourism Office would be preferable, to differentiate it from such bodies as Devon County Agricultural Association which are not part of the County Council. Also I will be inclined to make more entries under the publishing body than BNB, who frequently prefer title entry.

With regard to the detail of description one should bear in mind that a definitive entry is being compiled. It is for an item which is to be in the collection permanently, it may not be recorded in other databases, and it may be on closed access. The level of information given therefore should be as full as is reasonably possible. Besides, it should be a matter of professional pride that if the lending library, thanks to BLAISE (British Library Automated Information Service, 1977), can provide an entry for the latest Barbara Cartland complete with imprint, pagination, height to the nearest centimetre, and ISBN then the local studies librarian should be able to do the same for a local guidebook that will remain in the collection long after the good lady’s deathless prose has been disposed of via the library book sale or council incinerator.

However if AACR2 should form the principles on which the entry is constructed I have my doubts whether MARC (Machine Readable Cataloguing, 1968), at any rate in its full dress form, is an appropriate format for the computer input of the entry. The number and complexity of field indicators is too great for a system that would be largely used for material other than books, a case of a sledgehammer to crack a nut, and for a system that would need to be usable by non-professional staff with a minimum of training. In fact the entry for any category of materials can be broken into a relatively few elements that might require separate access for keyword searches or print out:

1.	Record control number.	This can double is an accession number or can use generally accepted reference numbers such as ISBN or BNB number.
2.	Special instructions.	This can include codes to facilitate printouts in special booklists, such as a regional bibliography. It could be combined with section 3.
3.	Category of record.	This assigns the record to a group such as monographs, periodical articles illustrations, maps.
4.	Description of item.	This includes title, author statement, edition statement.
5.	Details of production.	Place of publication, publisher, date. In the case of periodical articles periodical and volume.
6.	Physical description.	Pagination, illustrations, maps etc, size.
7.	Notes	These can be for public or administrative use and can be in free text structured or coded form.
8.	Author headings.	Also headings for joint author, editor, illustrator
9.	Subject headings.	In structured form giving Place, subject, name, aspects, date.
10.	Copy data.	Library, shelf location, stock number, copy specific notes (donor etc)

Such a limited range of distinct fields could be catered for by indicators consisting of a single word or number but it may be that MARC can be boiled down sufficiently to provide the suitably simple format. Whatever concessions to MARC local studies librarians might have to make in the question of the format of entries there is one thing that I feel that they must stand firm on, adequate subject access.

There has been a great diversity of approach to subject indexing local collections. At one extreme there are very broad subject groups: Architecture, Geology, History, perhaps a dozen or two all-purpose terms, redolent of a gentleman's library of the 18th century and sufficient for a few shelves with a couple of hundred items, but hopeless for the 20th century geologist who requires information on mineralization in the metamorphic aureole of the Dartmoor granite batholith, the architect who is studying the design of Victorian public buildings, or the historian interested in Civil War campaigns. At the other extreme are detailed analytical indexes to individual works or collections of works providing a plethora of one-line references to very specific items but submerging any meaty chunks of information beneath an ocean of gravy. A rich quarry for the antiquarian and genealogist perhaps but effectively dismembering bodies of knowledge that previous researchers have painstakingly gathered together.

Sometimes the two extremes can be found under the same roof. The librarian realizing that the broad arrangement adopted when the collection was set up is now insufficient to cope with its incremental size produces a range of analytical indexes or seizes upon those produced by local antiquaries and builds upon them.

If he is scientifically inclined, he will devise his very own classifications scheme. It may just be Dewey with the -0942 topographical subdivision lopped off the end, and the 900s re-cast so that 951 is not China but Chester or Cheltenham. He may be more ambitious and evolve his own classes and subclasses. A two-figure base is popular to avoid confusion with the three-figure base of Dewey. Guildhall library and the scheme of J. L. Hobbs are examples of this. The advent of faceted schemes widens our possibilities still further. I have in my time being responsible for a faceted scheme for a horological collection and a modification of Dewey. Whether the faceted scheme is understood and utilized since my departure from that particular special collection I have not dared to inquire and the modified Dewey cannot cope with many of the subjects represented in the local collection without a great deal of twisting and lateral thinking.

In my more cynical moments I feel that the only obvious claim to professionalism that librarians may have in the eyes of the general public derives from their ability to hide a book by the application of a long and complex number to its spine. Those of us weaned on Dewey know the feeling of being cast a drift without a paddle when let loose in a library classified by the Universal Decimal Classification let alone the Library of Congress scheme or Bliss. Perhaps it is the idiosyncrasies of our wide variety of classification schemes that have caused local collections to be left to one side when there has been discussion of computerisation.

Are we afraid the without these hieroglyphics it will all look too easy? We all speak the same language. Why then devote precious time converting English into a variety of incomprehensible symbols which in many cases only approximate to the actual subject content of the item indexed.

The more perceptive among you may gather from these remarks that I do not see classification schemes as the primary means of subject access in local studies indexes. This is not to say that I see no use for some form of subject grouping of items on the open shelves of local studies collections. The first three figures of Dewey are all the detail that is needed to specify a subject, after all it is a scheme familiar to most library users the first three or so letters of the place name will identify the locality and the date of publication assigns a specific reference number to the individual item. Thus a label bearing 361/EXE/1906 could indicate a copy of the Charity Commission’s report on the endowed charities of the city of Exeter published in the year 1906. For histories, descriptions and guidebooks of various communities, where the place is the subject, the first three or so letters of the place name followed by the date to which the item refers should be all that is needed. Thus the label bearing the notation DEV/1954 would indicate a history of Devon to 1954 or a guidebook to Devon published in 1954.

There is no need to make the notation used to arrange books in a reasonably useful order on the shelves double up as a means of indicating their subject matter in library catalogues. Many local studies items are on very specific subjects and demand very specific subject headings. These can only be provided by structured verbal headings from which individual terms can be permuted in printouts or which can be searched online by keywords.

Some of you may have shared my frisson of delight when, a couple of years ago, the PRECIS (Preserved Context Indexing System, 1974) strings in British National Bibliography entries were recast to bring the place element more consistently to the beginning. After all, place is in many ways paramount in a local studies collection. The PRECIS strings themselves are detailed and have a grammatical structure which makes clear the relationship of terms within the headings. They reflect the subject content of the item in a way which Dewey on its own does not.

While I feel that a structured subject heading such as the PRECIS strings provides the answer for subject access, I am not without criticisms. Some of the headings as employed by BNB are precise to the point of pedantry. I cherished such things as Livestock: Horses. Riding for Horseriding or Fiction in English. Special subjects: Vertebrates for Animal stories, but admit that compromises between natural language and indexing language are often difficult to arrive at. Rather more of a drawback is the system of shunting the terms to produce two-tiered headings which can appear confusing to the inexperienced user as well as involving a series of indexing operations which could be too complex for non-professional staff.

Keyword subject indexing involves the selection of one or more subject terms to ascertain how frequently and where they occur in a database, either alone or in combination. In the manual indexing system this approach can be catered for to some extent by extracting successive subject terms from a subject string and placing them in turn above a unit entry card which repeats the whole heading in its original order together with the catalogue entry for the item itself. I have been experimenting with this form of subject indexing on a selection of local periodicals by making a full descriptive entry on a card together with a full subject heading. This provides the unit entry which is then copied the appropriate number of times, each term being added at the head of each unit entry.

The form of heading I have devised is made up of a series of elements arranged in an order which makes each syntactically dependent on the proceeding term. While the precise order can vary, the normal order of the elements is given in the table below.

[In 1983 I had not yet discovered Kipling’s poem, which provides the fields for the subject string currently in use for the Devon bibliography.

I keep six honest serving-men
   (They taught me all I knew);
Their names are What and Why and When 
   And How and Where and Who.

This suggests a series of linked elements or fields:

Element 1983	Kipling poem	Field 1985+	Details
Context.	Where?	Region.	Westcountry, Devon, Cornwall, Somerset, Dorset.
Place.	Where exactly?	Place.	Community (town, parish) or major geographical feature.
Entity	What?	Subject.	Concrete rather than abstract: Prefer Mines to Mining.
Name	Who? Which?	Name.	Personal name. Family name. Building name. Organisation name. Ship name etc.
Part	How?	Aspect. Showing order in which individual elements are applied.	Towers. Rolling stock. Livestock etc
Property
Action			Planning. Conservation. Construction etc.
Agent			By … Role of … Public consultation.
Relationship	Why?		Related to … Compared to …
Viewpoint			Marxist viewpoint. Conspiracy theories etc.
Form			Reports, surveys. Policies and programmes. Indexes etc.
Date	When?	Datespan.	Of content of document, not always of publication.

When each term is repeated in turn at the front of the heading it relates immediately to the place element only, which is not always particularly helpful for terms which appear well down in the heading but more sophistication is difficult in a manual index without complex operations to commute the terms more flexibly. However this system, very economic in time for an interim manual application, is easily transferable to a computer application. I have used it to tackle the frequently complex and highly specific subjects in periodical articles and, hopeful that the same method could have a wider application, I have also been employing it for indexing illustrations. This has immediate implications for the choice of preferred subject terms. For illustrations, as for directory type information, the concrete term is normally preferable. One may have a book on banking but the illustration will be of a bank, or possibly a portrait of a banker. Directories have lists of banks not of bankings. This is no great problem for a local studies collection, where most information is not about the subject in the abstract but about its practice in specific places. The history of banking in Exeter is a history of its banks. Even if the theory of banking is discussed, it could be accommodated by some such combination as Banks. Operations. One cannot completely veto the inclusion of the abstract term but it should only be necessary to employ it relatively rarely. More common is the choice between product and producers, clocks versus clockmakers, beer versus brewers. The use of both terms cannot be avoided and one cannot hope to combine them in the much more cumbersome form as Beer and brewers. The producer and the product are different and are subjects that can be combined in different ways: one can wind up a clock but one cannot wind up a clockmaker.

So we have arrived at the concept of a local studies collection with a very specific subject approach possible through verbal subject index, searchable on cards on microfiche printouts or by keyword on an online database. This is all very well for the researcher who wants information on mineralization in the metamorphic aureole of the Dartmoor granite batholith, but what about the person who wants a list of documents on geology or rocks? How would the library be able at a press of the proverbial button to produce a listing of works on broader themes such as geology, transport, sport or health services? There is nothing inherently obvious in the combination of letters that makes up the word granite for the computer to realize that it is a type of rock or in the letters contained in the word mineralization to reveal it is an aspect of geology.

A thesaurus is necessary to convert these random combination of letters into a language that the computer can understand. It is not a classification scheme in the traditional sense that we are looking for, nor perhaps even a thesaurus in the traditional sense. It should be thought of rather as a dictionary that defines each term for the computer. The general public and in many instances even the indexer would not need to worry about its existence once it has been developed. Its development however requires a long series of careful decisions. [The world of search engines has moved on since 1983 but the somewhat laboured explanation which follows is retained as a period piece from the early days of computer applications.]

Let us take an example of how such an index language would work. A planner wants technical reports on the provision of parking facilities for football stadiums in the Plymouth area. The search on the key terms produces nothing and he wishes to widen the search. He is faced with the problem of searching the more general term Sports facilities and still finding nothing, although material may be hidden under the terms Cricket pitches, coded ABCE, Swimming pools, coded ABCF, or a score of other locations. If the term Football stadiums could be matched by the computer against a vocabulary given the code ABCD, that located it among Sports facilities, coded ABC, which in turn is an example of Leisure facilities, coded AB, which in turn falls within an overall grouping of Social institutions coded A, then the searcher could widen the search to any level required by the simple process of truncating the code and instructing the computer search for all terms covered by the truncated code thus ABC* would automatically pick up not simply Sports facilities, coded ABC but also Cricket grounds, coded ABCB etc. If required, a similar procedure could be operated on the term Parking facilities, coded EFGH, an aspect of Road transport, coded EFG, itself a form of Transport, coded EF, which in turn falls into the overall grouping of Communications, coded E. The geographical area of search could be widened too Plymouth coded IJKL is situated in West Devon, coded IJK, Devon as a whole is coded IJ and the Westcountry is coded I.

In a way what we are looking for is a type of faceted classification scheme (as pioneered by S. R. Ranganathan) but it is intended to enhance verbal subject access and not to replace it. Nor need it be static and difficult to amend in the way that the schedules of Dewey are. As such a dictionary can be stored separately from the computer in the computer from the database itself, the regrouping or insertion of new terms in the dictionary need to have no more effect on the subject strings of the entries than the addition of an extra “see also” card in the subject catalogue or the publication of a new English dictionary would on the works of Sabine Baring-Gould. In fact, the ability to refer to such a dictionary could well ease the problems of vocabulary control within an individual index or be able to cope with variations in terms from one database to another. It could, if properly constructed, even correct an indexer who are keyed in an unacceptable or imprecise term.

Such a thesaurus would have a structure the bears little relationship to the traditional disciplines which Dewey felt obliged to honour back in 1876. Indeed it can be very difficult to avoid the subjective or traditional connotations of many terms when attempting to assign them to their correct groups. For example the term Old persons to many people is automatically a social problem and could thus be assigned to the list of terms connected with welfare work: Almshouses, Institutional care. Not only is this an insult to many capable old people but, even worse, it is an inaccurate classification. Only in combination with terms such as Care homes or Hospitals should Old people become social problems for the indexer. In isolation they belong to Social groups and that is the wider term that they should be sought under. Indeed one could go further and factor the terms into Old a temporal quality and Person a representative of humankind. The question of how far one should factor terms is a major problem in drawing up this type of thesaurus. A term like Thermometers can be searched as a keyword. But the thesaurus might factor this as temperature + measuring + instruments. Thus the research into the history of instrument makers in a locality could access entries for the Barometer makers, Thermometer makers and Clock makers without needing to know in detail how the thesaurus broke each term down. The indexer too need not know that when he keys in the term Clock makers the thesaurus breaks it down as Time + Measurement + Instruments + Manufacturers, although he might be informed that he should reconsider the use of a term such as Timepieces.

Some of the refinements offered by a dictionary of this type are possible in a manual system by making a copy of it available for direct use by the searcher or by the insertion of cross-reference cards in the index which relate each term both to both higher and lower terms in the hierarchy. The searcher is thus able to extend or narrow his search although much more laboriously than within an online system.

I have given some thought to the form that such as thesaurus should take for local studies databases and have gone so far as to attempt to isolate and group terms. A main problem is the level of factoring and the effect that this has from the number and contents of groups of terms. Also difficult is the adaptation of the abstract groupings such as occur in general thesauruses to the practical needs of such a wide ranging field as local studies. Roget’s six main groupings of Abstract relations, Space, Matter, Intellect, Volition and Affections do not do much for the user of the local studies collection. The three main groups as far as the user concerned are probably Natural environment, Society and Locality.

[I had toyed with the idea of a pronounceable notation … ] with the nagging thought at the back of my mind that I'm maybe the latest in a line of lost causes which extends back at least a millennium to Amara Sinha who produced a classified vocabulary of Sanskrit in the 4^th century. It also includes George Dalgarno who in 1661 produced Ars signorum, the first pronounceable classification scheme which he provided with a grammar to make it into a philosophical language. As a sample of his work “In the beginning God created the heaven and the earth” would translate as “Dan semu, Sava samesa Nam tηn Nom”. Both heaven, Nam, and earth, Nom, belong to the category N, Natural objects. To my mind it is a formidable achievement and successors from Bishop John Wilkins with An essay towards a real character, and a philosophical language, published in 1668, a system of philosophically derived symbols, through the plethora of artificial languages Volapük, Esperanto, Ido, Interlingua, Novial, right down to Blissmybols or Semantography, published by Charles K. Bliss in 1949, all have less to offer to the deviser of this type of vocabulary than does Dalgarno himself.

[My immediate inspiration in 1983 was the London Education Classification, a library classification and indexing thesaurus used at the University College London Institute of Education, devised by Douglas and Joy Foskett in 1963, and revised in 1974. It is a faceted classification, inspired by the work of S. R. Ranganathan and of the Classification Research Group. Soon after this 1983 talk I abandoned the idea of a pronounceable notation. The alternating 21 consonants and five vowels were more restrictive in drawing up schedules than the regularity of 0-9, and, after all, when the coding is meant to be hidden, what is the point of making it pronounceable? An interesting side-alley for those linguistically inclined to wander down though.]

[… By whatever route we arrive there,] we should all be thinking of ways of utilizing computer systems and this is the direction in which much experience and not a little thought on the problems of local studies indexing has propelled me. Computerization offers us the possibility of making our indexes more flexible and more specific without making them less easy to use. It also offers us the possibility of hiding the often cumbersome means by which we seek to arrange information by subject. This in no way diminishes our professionalism; we may no longer earn the admiration of users for the way we can solve the self-imposed problems surrounding classification schemes but we shall more often earn their thanks for leading them swiftly to the specific piece of information they require. Computerization also offers us the opportunity of making a clean start and, if we have the will, a unified approach throughout the country to the organization of our local collections. I think it is an opportunity we should not miss.

[How naive that all sounds now to an old bibliofool like me!]