SOME COMMON PITFALLS IN THE DESIGN OF ONTOLOGY-DRIVEN INFORMATION SYSTEMS

Ontologies are the mean proposed by the Semantic Web to manage the knowledge of a system. In the software development industry, however, there is another de facto standard set of modeling techniques, methodologies and tools. Reasons for that issue might include technological challenges still to be solved, but also the need of a deep understanding of ontology modeling and the role that the ontology itself should play. In this paper we show that: 1) building an ontology is an error-prone task that might hide unexpected difficulties even when the ontology is apparently complete and correct, and 2) there is a tendency to misunderstand the role of ontologies when compared to other technologies such as relational databases, forgetting the benefits and strengths of combining both. For such a task, we develop and analyze a case study information system to recommend recipes and menus. We detect, classify and propose solutions to unexpected design pitfalls.