UK Union Catalogue of Chinese Books

BACKGROUND

In the UK, as elsewhere, the development of library systems to handle CJK (Chinese/Japanese/Korean) and other non-roman scripts has been very slow to materialise — only recently have two libraries (Durham University Library and the Library of the School of Oriental and African Studies, London) installed a multi-script system (INNOPAC). COPAC, the national union catalogue of the libraries of UK higher educational establishments, has stated its intention to implement a multi-script capability, but has not yet done so. Whether on the local or national level, the development and delivery of multi-script catalogues in all European countries has always been regarded as desirable, but never prioritised. But for readers of CJK materials, no less than for CJK librarians, original script catalogues are essential, owing to the gross ambiguities of romanisation in these three scripts uniquely (the situation is quite different for non-roman scripts that are alphabetic).

RLIN were the first to offer original script cataloguing, having implemented their CJK facility in 1984. In February 1986, they loaned a terminal to the British Library for local assessment, and in the course of that year it was demonstrated to East Asian librarians in the UK. By October 1986, the BL's assessment had been made public, and the system found disfavour on account of its expense, inflexibility, and the fact that it did not use Hanyu Pinyin romanisation (— the BL had switched to Pinyin in 1966, and was probably the first western library to do so). The input system (through a complex, character-component keyboard) was found to be a particular problem, and since using it with any fluency required considerable training and experience, it was found unsuitable for local conditions, where CJK libraries are not nearly so well-staffed as in the United States.

None of the features of RLIN that gave rise to these objections was changed until June 1996, when the RLIN Terminal for Windows was released. This did away with the complex keyboard and provided a Pinyin input system, but the romanisation standard for the data continued to be Wade-Giles. Meanwhile, the need for automated original script cataloguing throughout Europe was becoming acute — discussion of it had been on the EASL (European Association of Sinological Librarians) agenda since its second conference in 1982 — and it was noted that in the CJK countries themselves, radically different approaches were being taken, and that a number of alternatives to RLIN were now available.

The first to be taken up was NACSIS (National Center for Science Information Systems), now known as NII (National Institute of Informatics), which in 1991 was offered free of charge to a consortium of UK libraries with the aim of facilitating the cataloguing of Japanese material and thus contributing towards the formation of a worldwide union catalogue of Japanese books. In an area where disagreements among specialists are profound, this project has been outstanding for the unanimity and single-mindedness of its participants: all major UK research libraries are taking part, and all have made substantial progress with their retrospective catalogue conversion, some have even completed it. The result is a large and growing UK Japanese Union Catalogue, accessible to readers from versions in both Oxford and Cambridge.

For China, the situation was much more difficult, as there was (and still is) no equivalent of NACSIS, much less an invitation to participate free of charge. A way forward was presented at the EASL conference in Leiden in September 1990, when there was a demonstration of allegro, the library software developed by Bernhard Eversberg at Braunschweig Technical University. The software not only fulfilled all the standard requirements of a library system at a fraction of the cost, but was found to work perfectly in conjunction with Chinese front-end processors, accepting original script data in exactly the same way as romanisation. During the course of 1991, allegro was configured for use at the Bodleian Library (Oxford), and implemented for both librarians and readers in January 1992. The model was replicated at the Brotherton Library (Leeds) in February 1993, and at the British Library in November 1993. In Oxford and Leeds, retrospective cataloguing was given considerable impetus by three-year projects under HEFCE's recently completed "Non-Formula Funding of Specialised Research Collections in the Humanities Initiative".

Other libraries found Oxford's allegro model unacceptable for a number of reasons, among them its use of a non-MARC standard, and the fact that it was a separate database, not integrated into the main library system. Another consideration was the large size of the RLIN database, and the possibility that rapid retrospective conversion might be achieved through copy-cataloguing. Accordingly, Durham implemented Chinese original script cataloguing in its new INNOPAC system in March 1998, followed by SOAS in October 1999, using UK MARC and US MARC respectively, and RLIN as a source for derived records. Cambridge is developing its own in-house solution, based on the automatic matching of its existing romanised records with Peking Library's China MARC database, which it is reasonable to assume will become the largest source of Chinese bibliographical data in the near future.

The use of three different library systems (and four formats) in UK sinological libraries would not itself prevent the straightforward provision of a single access point to the databases (through the Z39.50 protocol, for example), but the situation is further complicated by the adoption of radically different standards in character coding and romanisation — although there is no dispute over the use of Hanyu Pinyin, opinions vary as to whether there should be syllable aggregation. The differences are set out in the following table.

However, there is at least total agreement in the two most fundamental areas: the observance of cataloguing rules, and the use of a format that makes bibliographical distinctions (rather than presenting data as an unstructured text file, for example). This makes it possible to produce order from a chaos that is more apparent than real. The following pages will define the project, and explain how it will be done.

TOP