The importance of the descriptive catalogue of manuscripts as a bibliographic tool is currently being reaffirmed by the interest shown by various institutions in creating such catalogues on-line. The advantages of an electronic format for such a tool--accessibility, searchability, ease of revision and expansion--are too widely recognized to need explication here. Instead, this paper concentrates on some of the difficulties involved in the attempt to exploit these advantages without sacrificing the usefulness of the mature traditional printed format, particularly in the case of non-Western texts; and on the solutions employed in designing one such project, an electronic descriptive catalogue of some 5000 manuscripts from South and West Asia.
The task of descriptive cataloguing itself, irrespective of the format of the finished product, exhibits a new set of problems to those cataloguers who venture outside the European textual tradition. In the first place, the greater chronological extent of the chirographic tradition in most of Asia, as compared to that in most of Europe (where scribal copying of texts began later and was sooner replaced by printing) means that there are simply more manuscripts from the non-Western world, by some orders of magnitude. Related to this is the comparative lack of bibliographic control for non-Western manuscripts, where the relation among author, text, and manuscript is generally much more obscure than for European ones. Thus any catalogue of such manuscripts must undertake to be also a catalogue of the texts themselves, rather than merely concentrating on the physical characteristics of manuscript instances of known works.
The same uncertainty permeates almost all other aspects of Asian manuscriptology. Even the subject classifications of known works are not firmly established in all their divisions: although a sufficiently minute classification scheme may exist in the work's indigenous tradition, such schemes are in many cases not yet usefully integrated into Western bibliography. The features of the physical manuscripts are no more tractable than those of their more abstract contents. Since it is neither feasible nor sensible to refrain from cataloguing non-Western manuscripts until all these varied issues of bibliographic and codicological control have been resolved (and indeed, in all probability it is only the knowledge gained through cataloguing attempts that will make it possible to resolve them), the author of such a catalogue must for the present rest satisfied with ensuring that its structure and presentation convey as much information as possible, without giving misleading impressions of certainty where none exists.
The foregoing brief and simplified description of some challenges peculiar (either in kind or in degree) to making a catalogue of non-Western manuscripts should serve to make it apparent that the electronic finding aid, implemented as or at least based upon a fielded database, is not in all respects the ideal model for producing such a catalogue in electronic form. The searchable fielded database works best with data that conforms well to a highly organized and unambiguous structure, and that can easily be segmented and then reconstructed in an automated process. The varying confidence levels of the information comprised by a catalogue of non-Western manuscripts are not as well suited to this format. Finally, a descriptive catalogue in electronic form should have a structure that is easily converted to that of a conventional printed catalogue, with the standard textual finding aids that electronic search capabilities attempt to mimic and extend.
These are the considerations that shaped the design of the bio-bibliographical electronic cataloguing project described in this paper, a project now being developed by the American Committee for South Asian Manuscripts (ACSAM). ACSAM will publish over the next few years an on-line version of a Union Descriptive Catalogue containing detailed information about thousands of South and West Asian manuscripts in North America and the texts contained in them. The prototype of the Union Descriptive Catalogue reflects some approaches differing from the design of most traditional electronic finding aids; they were adopted to lessen the tensions described above between the binary logic of computer operations and the ambiguity inherent in describing works and manuscripts outside the European tradition. The chief of these approaches is the abandonment of a fielded database structure in favor of SGML.
The catalogue's user interface also requires some deviations from that of a more straightforward electronic finding aid, in order to accomodate the unusual nature of the data. Although non-roman scripts will all be represented in roman transliteration, the transliterated characters must include a complete set of diacritical marks in order to provide unambiguous standard scholarly character equivalents for each script. At the same time, the screen representation of these characters must be as little dependent as possible on the advanced capabilities of current hardware and software, since many users will not have access to the latest technology or anything near it. Finally, the search engine will have a customized query manager to augment its existing capabilities of searching for free text and/or by element or attribute type, or structural characteristic. The query manager will employ various "fuzzy matching" techniques to optimize the usefulness of searches whose typical ambiguities will include variant forms or vocalizations of names, comparisons of precise dates or chronological ranges with vague or unknown ones, identification of well-known texts by the European version(s) of their titles, and so forth. The goal of the design as a whole is to use as much as possible of the power of computing capabilities while simultaneously doing no violence to the somewhat amorphous structure of the data.