"Every service or dataset in a distributed system makes claims about 'what it is' or 'what this item is' or 'what list this term comes from'. This means that "Controlled Vocabularies", or more generally classifiers are at the base of the interoperability stack. Every other service or dataset needs to be able to refer to items from lists, or the entire lists. In some cases such vocabularies can be used to populate menus. In other cases it is merely a matter of ensuring that comparisons are between like things. In the most sophisticated cases 'semantic mediation' relies on either reference vocabularies or at least requires that the definitions being compared are available through the same networks.<br />
The emergence of Semantic Web technologies has provided some powerful tools for formalizing definitions, vocabularies, and ontologies, in forms that also support reasoning and inferencing. In particular, the Simple Knowledge Organization System (SKOS) is based on a vocabulary/thesaurus model, and was designed primarily to formalize existing vocabularies using the semantic web tools, and also smooth the transition towards the richer logic-based tools from ontology modeling.<br />
However, so far the standards only provide very low-level interfaces to vocabulary data. For example, the basic query language for RDF data is SPARQL. While this is very powerful, it is at a low level similar to the relational database query language SQL. In the same way that standard database users would never interact with databases using SQL, standard vocabulary users should be provided with an appropriate interface which is based on the structure of vocabulary data, not generic RDF queries.<br />
This is the goal of the SISS Vocabulary Services project: to define a standard interface through which standard vocabularies can be provided to web users. The interface is primarily intended for machine-machine use, so that data providers can use HTTP links to vocabularies, data applications can be configured with standard terminology, and data clients can retrieve definitions or verify the existence of items claimed to be in particular vocabularies. Human interface in the form of web pages and forms are also required, but they are secondary to the data-oriented API.<br />
Note that SISSvoc does not address the issue of vocabulary governance. Its goal is limited to providing a user-friendly web interface to a vocabulary already available from a SPARQL endpoint, [currently read-only]. Vocabulary governance is a tricky thing to generalize since it involves both technical and social aspects. For the SISSvoc deployments at CSIRO we manage the vocabulary content as RDF files in a VCS. Vocabularies are deployed as a file and to an RDF triple-store with a SPARQL endpoint, triggered by creation of an SVN tag."
|search in BARTOC