Applications of A Categorical Framework for Conceptual Data Modeling

A.H.M. ter Hofstede,E. Lippe,T.P. van der Weide
DOI: https://doi.org/10.1007/s002360050112
1997-01-01
Acta Informatica
Abstract:For successful information systems development, conceptual data modeling is essential. Nowadays a plethora of techniques for conceptual data modeling exist. Many of these techniques lack a formal foundation and a lot of theory, e.g. concerning updates or schema transformations, is highly data model specific. As such there is a need for a unifying formal framework providing a sufficiently high level of abstraction. In this paper, focus is on the applications of such a framework defined in category theory. Well-known conceptual data modeling concepts, such as relationship types, generalization, specialization, and collection types are defined from a categorical point of view in this framework and an essential advantage is its "configurable semantics". Features such as null values, uncertainty, and temporal behavior can be added by selecting appropriate instance categories. The addition of these features usually requires a complete redesign of the formalization in traditional set-based approaches to semantics. Applications of the framework in the context of schema transformations and improved automated modeling support are discussed.
What problem does this paper attempt to address?