Abstract

Recently, ontologies emerge in various domains and different communities. For instance, in the database area, they are used to facilitate data exchange and integration, in the Semantic Web field, to describe terms, retrieve information and interconnect Web services. Several ontology languages and formalisms are given . Due to the ontology evolution, various applications manipulate a huge amount of ontological data . Consequently, they cannot fit in the main memory. To offer efficient storage and querying mechanisms of these data, academicians and industrials proposed database solutions storing in the same repository data and the ontology describing their senses. Such database is called ontology-based databases . Despite this trend, no design methodology adapted to OBDB exists contrary to traditional databases, where two main methodologies inspired from: Codd-Armstrong and Chen research efforts are proposed. Developing a design methodology for OBDB constitutes a crucial challenge for the success of OBDB applications. Similarities between conceptual models and ontologies recently identified motivate us to propose an OBDB design methodology inspired from the existing ones. To do so, we first describe in details these similarities and differences. Secondly, we propose an ontology-driven methodology to derive a normalized logical schema of a database from domain ontology whose each class contains a set of FDs. Finally, a prototype implementing our methodology is presented