Decidable logics for the Semantic Web

The notion of Semantic Web involves the publication of data in machine-readable format. Such data are intended to be enriched with so-called ontologies, which express information about the domain of the data, in particular about concepts and relationships among them. Reasoning tasks on data and ontologies are to perform effective information gathering and search. Datalog-based languages have been recently proposed for ontology reasoning and querying, in addition to the vast offer of languages based on Description Logics. In this talk we illustrate the main decidability paradigms that have inspired the Datalog+/- family, a family of Datalog-based ontology languages whose rules are Horn rules with the addition of existential quantification in the head. We show some complexity results for the problem of query answering under some relevant Datalog+/- languages, as well as some proof techniques.