Design considerations for small, special-system developmental databases Jamie A. Davies*, Andr´ e W. Br¨ andli†, David Hunter* and Pekka Nieminen‡ Small developmental databases, devoted to special systems and run by a normal research laboratory, differ qualitatively as well as quantitatively from larger databases described elsewhere in this volume. Resource limits set different optimum balances between computing elegance and ease of construction, while flexibility of design and semantics often assume greatest importance because specialist research data- bases are most useful when understanding of the biology is partial and still evolving. In this article we discuss the main considerations when building a small database— scope, access, data input, storage, searching and semantics — and illustrate them with examples of real working databases. Key words: database / development / ducted glands / HTML / internet / tooth / kidney ©1997 Academic Press Ltd THIS ARTICLE CONCENTRATES on databases that are devoted to special systems of particular interest to the developmental biology community. Such databases are of modest scope and are usually run as part of a biological research laboratory’s normal activity rather than by a team of computing specialists devoted to the enterprise. Although much smaller in scope than the large-scale projects described elsewhere in this vol- ume, the contents of special-system databases will often be much more detailed and may include preliminary, uncertain or controversial data (appro- priately flagged). The differences in aims, scope and resources available to special system databases when compared to large scale projects can result in quite different optimum strategies for their implementation. In this article, we shall discuss the main criteria that deter- mine the architecture of small databases and will then illustrate them by describing the designs chosen for the databases run by our laboratories, all of which deal with the development of specific vertebrate organs. To those contemplating the construction of a new database for their field, we offer the following three propositions to be borne in mind throughout the paragraphs that follow. Proposition 1. A misleading database is worse than no database at all. Proposition 2. A well-organized and maintained small database is more useful than a badly organized large one. Proposition 3. The need for a good database is greatest when global research effort into a field is at its most intense, at which time our understanding of a system will be changing most rapidly with the acquisi- tion of new data. Designing special system databases The design of a small special-systems database is influenced most strongly by the following considera- tions, which will be discussed in turn: • the aims and scope of database; • available resources; • access • methods for data entry • data organization and retrieval • database semantics • database maintenance Aims scope, resources and priorities The first and most obvious requirement of a database project is a set of well-defined objectives; the scope of the database, the type of data it will contain, the types of questions it will help to answer and, most impor- tantly, how it is supposed to help the research community. Typical issues to be decided in planning From the *Centre for Developmental Biology & Department of Anatomy, University of Edinburgh, Teviot Place, Edinburgh EH8 9AG, UK, †Institute of Cell Biology, Swiss Federal Institute of Technology, ETH H¨ on ggerberg CH -8093, Z ¨ urich, Switzerland and ‡Developmental Biology Programme, Institute of Dentistry and Institute of Biotechnology, Biocentre 1A, PO Box 56 FIN-000014, University of Helsinki, Finland seminars in CELL &DEVELOPMENTAL BIOLOGY, Vol 8, 1997: pp 519–525 ©1997 Academic Press Ltd 1084-9521/ 97/ 050519 +07 $25.00/ 0/ sr970176 519