Linked Data is a specific combination of formats and publication channels. RDF (Resource Description Framework) is the overarching linked data format framework, consisting of triples that each represent a ‘property’ or ‘field’ of an object description. The object described is always represented by a persistent identifier URI (PID-URI). The Library supports ARK (e.g. for items in the catalogue), Handle (image repository and scientific output) and DOI (for research datasets).
By using standard PID-URIs for metadata such as ‘author’, ‘subject’, etc., the objects described can be connected to objects, persons and concepts in other collections.
Many different ‘ontologies’ (data models) can be represented in RDF, including Bibframe, Dublin Core, Europeana Data Model (EDM), CIDOC-CRM and many others. RDF can also be presented in various syntax forms, such as RDF-XML, Turtle, N3 and JSON-LD.
The Library initially offers an EDM-based RDF format in Turtle and JSON-LD (see Formats).
The resulting RDF datasets can be offered through various channels, such as downloads, APIs and Sparql Endpoints. The Library collection RDF datasets are offered via the Triply-linked data platform, which features various tools for using the linked open data.
Individual object descriptions in RDF are made available on the web via content negotiation of the ARK PID-URIs for Accept-header MIME types application/json, application/ld+json and text/turtle. Content negotiation for other PID-URIs is available via the Triply API.