4. Computer systems

A further dimension to the questions discussed in the previous section, and a topic of increasing general relevance in any case, is the possibility of the involvement of mechanisation or computerisation in cataloguing. Computers are becoming more accessible year by year: the equipment itself is becoming cheaper, and there is an increasing choice of program packages for computer applications, including cataloguing. The only dimension lacking, perhaps, is a realistic appraisal of what computer usage mayor may not really involve.

Mechanisation has always been one way of reducing the drudgery of documentation. Even at a simple level, the use of stencils or photocopies to produce multiple copies, in whole or in part, of a catalogue entry can be one way of cutting down the clerical labour of generating several index entries. Full computerisation develops this advantage a stage further by removing the task of index compilation altogether. From a properly compiled catalogue entry, a good computer system will automatically generate all the indexes required with little or no further human effort. More, it will provide selective listings to suit special needs (it may, in fact, ultimately offer the facility of 'on line' dialogue between users and the catalogue data-base, virtually removing the need for indexes altogether)J it will also facilitate, through the technology of computer type-setting, the production of published catalogues.

Inevitably, for these advantages, there is a price to pay. Indeed, there may be three prices to pay, only one of which directly involves money. That one is, of course, the cost of the system. The others are restrictions on cataloguing procedures, and the changed nature of the actual task of cataloguing.

The first two are intimately connected: in crude terms, at present, the cheaper the cataloguing package, the more compromises an archive is likely to have to accept in its procedures. Since few computer packages have so far been tailored directly to sound archive needs, a sound archive is likely to have to use a package developed for other purposes or to develop its own. Access to an extant cataloguing package may quite possibly be among the more positive inducements for a new archive to conform to an institutional or national standard. The majority of such packages, however, are those designed for library purposes, and the involvement of a computer tends to intensify those difficulties already noted in cataloguing a sound archive to a library standard: for example, a computer system may be incapable of accepting more than a set number of 'responsibility' statements or a descriptive summary of adequate length, or of generating indexes from field recording details. On the other hand, the conversion, development or design of a more suitable package is likely to require access to skills which will not be available within the majority of sound archives, and can only be brought in from outside at some expense.

The third category of price to pay is the actual nature of work involved in cataloguing. It will be remembered that when it was claimed that a good computer system could automatically generate all necessary indexes, the significant qualification 'from a properly compiled catalogue' was added. A computer system is only as good as the information entered into it, and (unless time and money is to be spent on expensive error-checking procedures within the computer) the effect is that that information must be of the highest possible degree of accuracy and consistency, and on occasions of great apparent triviality as well. Human cataloguers will recognise that 'HMV' and 'H.M.V.' and 'His Master's Voice' are effectively synonyms, but to the majority of computer systems they would be totally distinct terms; similarly, human cataloguers may cope without effort with alphabetical filing conventions concerning the ordering of numbers and abbreviations, or the ignoring of opening definite or indefinite articles, but such details must be explained in some way to the computer. The price paid for trouble-free indexing, in other words, is more troublesome cataloguing and particularly more painstaking checking and proof-reading of catalogue information.

This is not intended to sound unduly pessimistic. Several archives are successfully using computers in their cataloguing, as the case studies show. It is, however, essential to be realistic, and especially to doubt the salesmanlike claims of anyone who alleges that computerisation will be cheap, or easy, or just like conventional procedures or (as is not unknown) all three simultaneously.