Résumé
We consider the application setting where a domain-specific knowledge base about Durum Wheat has been constructed by knowledge engineers who are not experts in the domain. This knowledge base is prone to inconsistencies and incompleteness. The goal of this work is to show how the state of the art knowledge representation formalism called Datalog± can be used to cope with such problems by (1) providing inconsistency-tolerant techniques to cope with inconsistency, and (2) providing an expressive logical language that allows representing incomplete knowledge.