Affiliated Organization

Indiana University, USA


There is a growing interest in developing a sound theory for the implementation of conceptual models. In the case of Entity-relationship (ER) diagrams, two conceptual models that are commonly used are local schemas and global schemas. Recent research has shown that global schemas that are generated from local schemas carry ambiguous information and thus fail to serve their purpose of communicating data requirements for organizations on a global scale. Eventually users are still forced to refer to local schemas to improve their understanding of data requirements. In this paper, we argue that these ambiguities arise due to a lack of formalisms related to conceptual models. We propose that conceptual models that are built according to a sound theoretical framework will give rise to ambiguous free schemas. Our study draws from the work on ontological rules derived the Bunge Wand and Weber (BWW) ontology and we focus on optional properties used in ER diagrams. Previous related research has already shown that conceptual models which follow the BWW ontological rules provide better understanding to analysts and users, but all of these works looked at a single schema only. Our work differs from these previous studies by looking at the effect of integrating different schemas together into a single global schema. We expect that by applying ontological rules to model information requirements in local schemas, resulting global schemas will be free of ambiguities and can be a suitable representation of user information needs at an organizational level without the need to refer to multiple local schemas.