Publication of data

Beyond the classic data exposure on this website <http://cocoon.huma-num.fr> with classic search and consultation tools, COCOON also offers an exposure that follows the principles of the "data web".

The "web of data" exposition

To expose Cocoon's resources following the principles of the "Web of data", metadata are translated into RDF syntax using the formal model of the Europeana project . This model allows, for example, to separate the description of a cultural object (for Cocoon, a speech recording) from the description of its representations (for Cocoon, the different formats and consultation modes of a recording).

As far as possible, the values present in the metadata are aligned with repositories. The main repositories are:

  • For authors: VIAF (Virtual International Authority File)
  • For geographical locations: Geonames
  • Pour languages: Lexvo
  • Pour subjects: RAMEAU (Répertoire d'autorité-matière encyclopédique et alphabétique unifié)
Some alignments have also been made with Dbpedia, TGN or LCSH.

The metadata structured in RDF are stored in a "triple-store" database manager. This database can be queried using the standard SPARQL protocol and query language.

To follow the principles of "linked open data" all Cocoon URIs are "de-referencable" and a redirection mechanism allows to deliver for the same URI its RDF representation (for machines) and its HTML representation (for humans).

Reférences

  • W3C page on semantic web standards
  • Europeana Data Model (EDM)
  • Example of a URI corresponding to a record: ESLO1: entretien 008
  • Query interface (with examples): SPARQL Endpoint