Engineering of KR-Based Support Systems for Conceptual Modelling & Analysis

Most of present intelligent knowledge management environments for conceptual modelling and analysis suuer, in our opinion, from mixing two representation levels: (1) a conceptual level, where domain-speciic concepts are represented (e.g., data and processes in Data-Flow Diagrams); (2) an epistemological level, where inferences are drawn from domain-independent linguistic construc-tors (e.g., concepts and roles in Description Logics). We propose an engineering approach to the development of new systems, where the two levels are separately represented and are either linked by the protodl methodology or by concept em-ulators. We exemplify our approach by modelling Data-Flow Diagrams in level (1) and by using a Description Logic with a Closed Terminology Assumption in level (2). The beneet of the proposed approach lies in decoupling the conceptual level from the underlying epistemological level in the overall architecture of support tools for the construction and the analysis of large domain models. This implies easy replacement of diierent inferential engines at the epistemological level, while retaining the same domain description at the conceptual level.