Eurovoc Conference - Mind the lexical Gap

On November 18-19 I participated in a workshop organized by the Publication office of EC in Luxembourg on the role of EUROVOC in underpinning the Semantic Web. (

The publication is office is strongly pushing EUROVOC as a cornerstone on an European Linked Open Data environment, focussing strongly on legal and political concepts. They were mentioning even a semantic web for legal information systems as pointed out by Martine Reicherts, the Director General of the publication office in her inauguration speech.

EUROVOC  has 6797 concepts at the moment and 6797 and preferred terms for every language. In 2007 they conducted a study about Thesaurus Management Software and     feasibility study about thesaurus ineroperability and mapping methodologies.  

In  March 2008 they issued  tender for a new thesaurus management system with a new data model and in December 2009a tender for Thesaurus alignment Environment.

The data model reflects perfectly what we did with AGROVOC in conceptualizing the thesaurus with the introduction of IDs/URIs for every concept. They encountered the same issues with the necessity to express also relations between labels and not only between concepts.  (Interesting issue in the Swedish translation of UNFCC, in which they are using the English abbreviations)


The tender for the ITM("Intelligent Topic Management) and the TMS (Alignnment) were assigned to Mondeca and a Belgian company.  The ITM is very similar to the concept workbench developed by our team with an OWL model behind and should be able to manage various KOS. The ITM exports SKOS/RDF into Drupal from which it is rendered to the public.

For the alignment work a consortium between Mondeca at INRIA-Grenoble has been formed. The alignment is now not taking a pivotal language but tries to exploit the multilinguality for a more efficient alignment.

The Participating thesauri were presented in a "speed dating" procedure.  AGROVOC was presented together with INSPIRE and GEMET, the two other vocabularies very well on their way to become linked open data.


The Cellar project

The cellar project is the Publication Office' s attempt to create a unique storage and disseminatio platform to access European law and publications. The existing silos and different access points will be integrated on a Fedora Commons Platform with a METS/FRBR data layer and integrated EUROVOC support. Metadata will  represented in RDF and stored in triple store.  Ontology base on FRBR defines the definition and data layer, SKOS to encode and express the authority files.



In the new guidelines will be a distinction between term and concepts

Important follow ups:

 - FAOs Geopolitical Ontology should be matched with AGROVOC, at EUROVOC there are plans for a European territorial classification with EUROSTAT together

 - The AGROVOC team at FAO should contact Jerome Euzenat for collaboration in Thesaurus alignment

- AGROVOC should establish all possible connections with EUROVOC when going online as LOD





Enhanced by Zemanta