Publishing the vocabularies is the first phase in the immediate SENESCHAL project and broader work. More to follow later but briefly flagging up immediate plans.
As good citizens of the LOD world, we want to make SKOS mapping connections both at the vocabulary level and more broadly connecting or collaborating with related projects. Currently there are a few mappings to Dewey linked data URIs, courtesy of a previous RCAHMW project. One obvious next step is publish SKOS mappings between the different national thesauri to each other and to the eventual uber thesaurus Phil has blogged about.
More broadly, making mappings between different language thesauri is also part of the ARIADNE archaeological infrastructure work we are involved in. This should also offer opportunities for other providers to make their vocabularies available as Linked Data.
To help with the mappings, Ceri is working on alignment, as part of a larger body of work on RESTful web services that operate over the vocabulary Linked Data and facilitate their use. We plan to build on previous work with SOAP based services for STAR and develop more complex higher level services attuned to the immediate SENESCHAL use cases as a first step.
Immediate applications of the services for SENESCHAL case studies include widgets for Browser based indexing applications.
Thinking of possible users unfamiliar with the heritagedata website or who may not even know that the vocabularies exist, how can we help them discover, assess and perhaps decide to use or even extend the vocabularies? Part of the eventual solution (as usual) is metadata – this time at the vocabulary level.
We have included some initial vocabulary metadata (see the Properties at the top level of any of the Linked Data vocabulary schemes) though it is something we want to come back to and extend. For example, we’d like to include some metadata on the broad coverage of the vocabularies. Eventually we envisage vocabulary registries through ARIADNE and related projects where users may go to search for relevant vocabularies.
For this metadata we have made a start by using some metadata elements from the DCMI KOS Application Profile (DC AP) effort led by Marcia Zeng. This presents a core and extended set of metadata elements and we will be adding to our initial set of elements through discussions with the vocabulary providers.
In fact, it’s not so simple necessarily as assigning metadata to a vocabulary (though this is where we are starting). We might want to distinguish between a particular publication manifestation (eg linked data) of a particular language expression of an underlying work. And yes the DC AP builds on the FRBR work of the library community and provides different metadata elements for the different parts of the model. We have not chosen to model at that level currently but are interested in any views on this.