The British Library
By Ulrich Pagel
The first concrete measures leading to the development of the
automated catalogue were taken in 1993. The database parameter were
defined by a number of library-internal decisions, influenced by
criteria such as budgetary restraints, format, contents, production
cost, and the final appearance of the printed product. From the
very beginning, it was agreed that the database should be designed
in such a way as to allow for traditional publishing in book format
and for electronic publication, possibly on CD ROM or the Internet.
The selection process of a suitable database application extended
over two months, during which six different commercially available
packages were tested. The following criteria were taken as
benchmarks in the selection process:
First, the application had to support unlimited field capabilities.
This was essential since the catalogue was to contain numerous
manuscripts requiring a very large number of different data fields
per entry. Second, it had to be structured so as to meet very
detailed record retrieval stipulations, allowing for searches on
any one field, in isolation and in combination with others. Third,
it had to be capable of supporting Tibetan, Chinese, and Indian
scripts, since script inclusion is indispensable for accurate and
unambiguous representation of the original titles. Fourth, it had
to possess picture capabilities, needed for image representation of
particularly interesting or valuable palaeographic exemplars.
Fifth, it had to be cross-platform compatible, in order to link the
British Library catalogue electronically with Tibetan cataloguing
projects taking place in France, Germany, and Japan. Sixth, it had
to support a relational file structure, imperative for efficient
database management, since searching for information in large
flatfile databases is prohibitively slow and inefficient. Seventh,
the application had to be extremely flexible, permitting changes in
the design at any stage to allow for the recording of unforeseen
manuscript peculiarities.
Very soon it became apparent that of the six database packages
tested only one would meet all seven requirements. This is 4th
Dimension developed by ACI France in the late 1980s. 4th Dimension
not only proved to be the most stable and versatile product, but
ACI also agreed to put one of its senior consultants at the
disposal of the British Library to assist in resolving structural
intricacies in the design of the Tibetan cataloguing database and
to fine-tune the user-interface in order to ease data input and to
systematize record retrieval.
Data input and retrieval
Apart from secure storage, the inputting and outputting of data
are, of course, the most important functions of databases in
general. Data input has to be intuitive and labour efficient while
data output should be flexible, detailed, and systematic. Data
retrieval needs to be flexible because most catalogue databases are
multi-functional. On the one hand, bibliographic databases are
research tools that serve as points of access to the texts
catalogued and provide ancillary bibliographic information which is
sufficiently detailed to be of use in academic investigations. On
the other hand, they have an important library-internal use, aiding
rapid text identification, housekeeping duties, project-related
calculations and so forth. Complete control over the inputting
interface and virtually unlimited data retrieval options provided
for by the source application (4th Dimension) allowed the British
Library to customize its Tibetan database into an extremely
resourceful cataloguing tool. For example, with the present file
structure, it is possible to search by author, title, folio number,
date, format, and script all at the same time, producing finely
grained retrieval results. Equally impressive is 4th Dimension's
handling of data retrieval procedures that can be very effectively
employed in stocktaking exercises: the total number of folios in
the collection (e.g., for microfilming purpose), average format
(e.g., for storage purposes), maximum and minimum number of folios
per volume (for boxing purposes), concordances, title sorting,
select bibliographies; this, and many more types of information can
now be retrieved within a matter of seconds. Another important
criterion in the database selection process was printing
capability, since most public enquiries are still conducted in
writing. 4th Dimension supports a very large number of printing
templates that may be used to obtain paper copies of retrieval
results. Once created, these templates may be used to specify
exactly the amount of data printed for each entry, depending on the
purpose to which the print-out is to be put. Flexibility in the
design of the printing templates means that their lay-out can be
adapted to match that of printed catalogues, approaching
publication quality.
During the past two years, there have been a number of European and
Japanese institutions which have expressed interest in the
catalogue structure of the British Library Tibetan database. For
the most part, these institutions plan to use it in the cataloguing
of their own Tibetica. However, there is nothing intrinsically
Tibetan about its current structure apart, perhaps, from its font
specifications. Similar database structures are used in the
International Dunhuang Project and in the cataloguing of the
British Library Khotanese and Burmese collections. The user
friendliness of the source application means that it is perfectly
feasibleþwithin a few days of basic programming activityþto adapt
its present structure and functionality to bibliographical material
in practically any language. Many of the fields do not need to be
changed at all, since they relate to general bibliographical
aspects such as title, author, physical condition, colophons,
dates, provenance, housekeeping, etc. In cases where changes do
have to be made, it usually sufficient to rename the fields in
question. Because 4th Dimension, and all databases created with 4th
Dimension, are cross-platform compatible (the Windows 95 version
was launched in October 1995), 4th Dimension databases can be run
on PCs and Apple Macintoshes. The database structure developed for
the British Library Tibetan collection is available free of charge
to bonafide scholars and institutions. The source application
required to drive the structure is commercially available.
All enquiries should be directed to
Dr. Ulrich Pagel
Curator (Tibetan Section)
The British Library, Oriental and India Office Collections
197 Blackfriars Road
London SE1 8NG
United Kingdom