Extending the conceptual graph approach for data conceptual modelling

摘要

Information Systems design methodologies include a phase called data conceptual modelling which corresponds to a true knowledge engineering activity. During this phase, relevant objects, their properties and proportions involving these objects are identified and described in a data conceptual schema. Designers often rely on unformal linguistic techniques in order to build data conceptual schemas. Semantics are generally specified by constraints on possible database populations and no meaning is associated with the proposition names, apart from intuitive meaning. The informal level of semantics is open to different interpretations, while the restriction of fact instances does not necessarily convey meaning to any observer. Alternatively, Sowa's conceptual graphs, while providing a linguistic level semantics are not ideal as a data conceptual schema language. This has led us to propose an extended conceptual graphs model. In this paper we discuss the data conceptual modelling process, we present the main features of E-R and NIAM models and we note some difficulties when using them for conceptual modelling. An extended conceptual graph model (ECG) which permits conceptual graphs to be used as a data conceptual schema language is introduced, and the equivalence between the ECG and E-R and NIAM models is discussed.