A framework for comparing conceptual models

Abstract

Conceptual models are a widely used mean for documenting software systems as well as describing organisational structures. The trend towards integrated and flexible information systems has encouraged research about the comparison of conceptual models. Current approaches on the identification of similarities between conceptual models often adopt an automation perspective only. In this paper we will unfold severe arguments that a fully automatic model comparison process is not feasible. Furthermore, we will show that only a semiautomatic process can perform the comparison of conceptual models at the semantic level. On this theoretical basis, we will develop a framework which identifies all necessary and sufficient components for comparing conceptual models. We will show that this framework includes all the requirements that a semi-automatic model comparison process must meet.