In-house usage scenarios

From the TC 372 Workshop Compendium

It often seems easier for libraries to share their catalogue with other libraries, for museums to share metadata with other museums, etc., than to connect different departments within a single institution. In-house metadata interoperability is still rarely achieved.

Why is it that most cinematheques collect various things in addition to moving images? It is because we tend to think of cinema as something that is embedded in a cultural context. How can we make our catalogues and other knowledge tools reflect this context?

In this example, we use a link-friendly library catalogue. Any catalogue record from the Deutsche Nationalbibliothek (DNB) can be retrieved via a URI that is agreed to remain constant throughout the lifetime of the institution. In this case, it is an HTTP URI that can be invoked directly from any Web browser: http://d-nb.info/1008010898. At present, the DNB catalogue is not as link-friendly as it could be. If it were capable of supplying the catalogue record in RDF, then standard software tools could immediately explore further paths, e.g. retrieving the authority record for the author (http://d-nb.info/gnd/142990698).

From http://commons.wikimedia.org/wiki/File:Reisestecker.jpg

How can we make a catalogue link-friendly? Define a Uniform Resource Identifier (URI) for each unit of information (e.g. catalogue record, or part thereof). Make sure that the URI can remain stable over a long period of time. Let a Web server respond to a HTTP URI request with the set of associated metadata (e.g. a catalogue record). Observe the rules of the game and use an agreed data model (i.e. RDF) and offer a choice of recommended encodings (XML, N3/Turtle, and something human-readable). Enrich your metadata with relationship statements. Use publicly available vocabularies for relationship types, and URIs for relationship targets. And that's about all.

References and Materials: In-house usage scenarios The Linked Data Book. Explains how to publish linked data. Can Bibliographic Data be Put Directly onto the Semantic Web? - Martha M. Yee (2009)