You have to ask yourself why you want to classify your objects. Is it so that you can retrieve items related to a particular enquiry, or is it so that you can arrange the objects in subject groups in store, or their records in a single sequence in a catalogue? It is only if you are arranging the physical objects themselves that you are restricted to giving each item a single classification number; like putting library books on shelves, each book may deal with several topics but it can go in only one place. A classification for museum storage purposes is in any case likely to be determined by other criteria: size, weight, environmental requirements, value, frequency of use, and so on. For information retrieval purposes, three approaches are possible, and as they complement each other they should all be provided for. 1. A classification scheme. This can be fairly broad, and can be determined by the local interests and needs of the museum. For example, objects can be classified by themes, areas of study, historical groupings, curatorial specialities, etc. Top levels of such a scheme might correspond to traditional major divisions of a museum: Art, Natural History, Technology, Ethnography, . . .. These can then be subdivided as necessary, but it is not worth using this approach to give specific identification of individual objects. The point is to provide useful groupings, of manageable size, so that people who want to browse around in an area - either in a gallery or in a catalogue - can find related items together. There is no reason why a single object cannot be assigned to two different classification groups. The example of an inscribed quilt could perfectly well be assigned both to the "documents" group and to the "textiles" group. The symbols you use for a classification scheme are just a way of making it easier to file items in the order you have decided will be useful. It would be worth your while talking to a friendly local librarian about the principles, especially how you can provide for future expansion and additions at any point. It's worth mentioning the words "faceted classification" too, as that may spark off a train of thought about structure. But keep it simple. 2. A thesaurus This is the main tool for specific retrieval of items. Its use assumes that you have a computer catalogue, even of a simple kind, so that you can combine terms at the time of searching. This avoids the need to construct compound indexing phrases, possibly with subheadings, as are used by Library of Congress Subject Headings; you just give each object all the terms which apply to it, and their order doesn't matter. A thesaurus has two main purposes: it ensures that you always use the same term to index the same type of object, and it points you to that term, at the times of indexing and of searching. It does this by a structure of relationships between terms which allows you to translate the words in an enquiry into the most appropriate terms to use to search the catalogue. 3. Free text indexing Many computer systems will allow you to search on the words used in a catalogue record. This is very helpful, but can give a false sense of completeness. Certainly the record for every object should be able to contain free text description and notes, but by their nature these cannot be guaranteed to be consistent in containing the same words for all objects of the same type. There are also no hints of related words which you might include in your search. Free text technique is best used to supplement a thesaurus, allowing you to pick out particular words which are more specific than the thesaurus provides for, characteristics peculiar to single objects, or ways of narrowing down a group once you have retrieved it in a structured search. The main reason why free text is widely used is that it is easy to implement - there is little intellectual effort in indexing at the input stage, because the burden of thinking of all possible search terms is loaded on to the enquirer. There is a great deal more that could be said about this, but I shall resist the temptation to pursue this pet enthusiasm of mine longer here. There are many books on both information retrieval principles in general and on museum documentation. Unfortunately there are no really good general thesauri or classification schemes for museums, which is why museum people have to think about principles rather than just being able to take something off the shelf and use it. A list of existing indexing vocabularies for museum objects will shortly be published by CIDOC (the Documentation Committee of ICOM, The International Council of Museums), and will be available at the CIDOC/MCN conference in Washington,DC, 28th August - 3rd September, where there will be substantial discussion of all these topics. Regards Leonard Will Treasurer, CIDOC Chair, CIDOC Working Party on Museum Libraries and Information Services -- Dr Leonard D Will Tel: +44 81 366 7386 Information Management Consultant Fax: +44 81 366 0916 27 Calshot Way, ENFIELD, Middlesex Email: [log in to unmask] EN2 7BQ, United Kingdom