Fitting reality into a database mold: Rules for data collection

摘要

Presents a typology of rules for database design through examination of one such set of rules, the Anglo-American Cataloguing Rules, Second Edition (AACR2). Analysis of the AACR2 reveals three clusters of rules: 1.(1) content rules;2.(2) format rules; and3.(3) data collection rules. The last cluster is of interest to database designers in general, and includes the following types of rules: 1.(1) rules for authorized sources of information;2.(2) rules for establishing entities, relationships and attributes;3.(3) rules for domains;4.(4) rules for occurrence; and5.(5) rules for cardinality. Further examination of these rule types shows particular instances which require specialized kinds of rules to resolve cases of absent, multiple, fuzzy or ambiguous data. Though useful for database design work as it stands, the typology could be enriched by addition of rule types from other databases and by further exploration of relationships between rule types and data characteristics.“An information system (e.g. data base) is a model of a small, finite subset of the real world.” [1]