Semantically unlocking database content through ontology-based mediation

  • Authors:
  • Pieter Verheyden;Jan De Bo;Robert Meersman

  • Affiliations:
  • Vrije Universiteit Brussel – STARLab, Brussels, Belgium;Vrije Universiteit Brussel – STARLab, Brussels, Belgium;Vrije Universiteit Brussel – STARLab, Brussels, Belgium

  • Venue:
  • SWDB'04 Proceedings of the Second international conference on Semantic Web and Databases
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

To make database content available via the internet, its intended shared meaning, i.e. an interpretation is required of the database (schema) symbols in terms of a so-called ontology. Such an ontology specifies not only concepts and their relationships in some language, but also includes the manner in which an application or service is permitted to make use of these concepts. Ontologies therefore also play a key role in making databases interoperate. The DOGMA approach to ontology engineering is specifically adapted to the classical model-theoretic view of (relational) databases. Noteably, it rigorously separates an ontology base of elementary lexical fact types called lexons, from the rules and constraints governing the concepts referred to by the lexons in the ontology base. These rules are reified in so-called ontological commitments of applications to the ontology base. In this paper we formalise and make precise the structure of this commitment layer by defining Ω-RIDL, a new type of so-called commitment language. Examples derived from its use in a non-trivial case study are provided. We illustrate how some of its key constructs, designed to specify mediators by mapping databases to an ontology base, can conveniently be reused in a conceptual query language, and report on its ongoing implementation.