Tedi: a platform for ontologisation of multilingual terminologies for the Semantic Web
Résumé
The vision of the Semantic Web is machine understandability for all data currently stored in web-based resources. Human and machine readable terminological resources, which follow the ISO standards on terminology in defining concepts as unique combinations of essential characteristics [ISO 1087-1], need to become computable and Semantic Web compliant. This paper, first, describes the theoretical approach and the tool-assisted method which underlies the turning of these terminologies into Semantic web compliant ontologies. Next, this paper presents Tedi (ontoTerminology editor), the platform developed for building multilingual terminologies which share the same formal domain ontology. Tedi allows to export these terminologies into OWL/RDF ontologies, JSON, and in a number of other formats, including multilingual electronic dictionaries of terms in the form of static and dynamic html. Tedi is based on a theory of concept dedicated to Terminology, the discipline which defines semantics as the relation between terms (natural language units with meaning specialized to a domain of knowledge) and concepts (units of thought whose meaning is formally expressed as a set of essential characteristics). Tedi stores the linguistic and the conceptual dimensions in two distinct, yet related systems. This formal theory which supplies the semantic onto-terminological layer needed for deeper data interpretability by machines is less contrived and far more intuitive to use. It empowers domain experts to build their own semantic multilingual terminological dictionaries without having to be aware of logical formalisms like description logics. Semantic content management systems are direly needed in the domain of ancient cultural heritage. The remainder of the paper will illustrate this particular point with a use case from the domain of ancient Greek dress terminology presented from the point of view of the user (domain expert).