Comparing data modelling frameworks using Chisholm's ontology.
Simon K. Milton, Edmund Kazmierczak, Chris Keen · European Conference on Information Systems · 1998
Data modelling frameworks consist of a series of constructs which are used to create an abstraction of the world. Ontologies consist of concepts through which reality can be described. The aim of this paper is to use an ontology as a basis for comparing and evaluating data modelling frameworks. We begin by establishing a method for comparing modelling frameworks using ontology, and then examine the role ChisholmOs ontology can play in such a comparison. We then perform a comparison of three semantic data models: the EntityRelationship Model, the NIAM model, and the Functional Data Model. In undertaking the comparison we introduce ChisholmOs ontology, which is a new ontology in the Information Systems literature. ChisholmOs ontology has given us new insights into the role of attributes and into the changing nature of relationships, although in this paper we do not discuss the latter. All the data modelling methods supported the ontology to varying extents, with the Functional Data Model offering the best support. However we found that all the models lacked the full generality of ChisholmOs ontology. The disagreement between current data models and ChisholmOs ontology provides insight into the requirements for a truly flexible data modelling technique.