Engineering Conceptual Data Models from Domain Ontologies: A Critical Evaluation

Ontology Vs. Conceptual Data Model

This section informally explores ontologies and CDMs, including their similarities and differences. The literature shows many definitions of ontologies with the most popular definition proposed by Gruber (1995) as “a formal, explicit specification of a shared conceptualization” (p. 907). In general terms, an ontology may be defined as expressing knowledge in a machine-readable form to permit a common understanding of domain knowledge, so knowledge can be exchanged between heterogeneous environments.

On the other hand, conceptual data models capture the meaning of information for modelling an application and offer means for organizing information for the purposes of understanding and communication(Mylopoulos, 1998). The major role of the CDM is to model the so-called universe of discourse (UoD), entities and relationships in relation to particular user requirements independent of implementation issues. Hirschheim, Klein, and Lyytinen (1995) define the Universe of Discourse in the information systems (IS) world as “the slice of reality to be modelled” (p. 58). Therefore, there are some similarities and differences between ontologies and CDMs. Both are represented by a modeling grammar with similar constructs, such as classes in ontologies that correspond to entity types in CDMs. Thus, the methodologies of developing both models have common activities (Fonseca & Martin, 2005). While ontologies and CDMs share common features, they have some differences. According to Guarino’s (1998) proposal of ontology-driven information systems, an ontology can be used at the development or run time of IS, whereas a CDM is a building block of the analysis and design process of an IS.