The ICONCLASS system has been available as LOD in SKOS form since circa 2015, but never had a SPARQL endpoint available for federated service queries.
Some periodic dumps in .nt format were done (but not of the full “expanded” set of ~ 1 million nodes) - and the URIs in “http://iconclass.org” form were used.
Retrieval of URIs under the convention http://iconclass.org/xyz.rdf where xyz a known Iconclass notation, returned the SKOS representation in RDF XML serialisation.
These were for example used by Europeana to resolve metadata for those collections using ICONCLASS.
Since 2023 we have made a first SPARQL endpoint available, and it has been used by the Dutch Netwerk Digitale Erfgoed for the Termennetwerk terminology service.
There have also been requests from some research groups (at Factgrid.de for example) to use the ICONCLASS endpoint in SERVICE calls to resolve prefLabels for given notations via SPARQL.
There are some shortcomings in the current implementation that need addressing.
The LOD version of ICONCLASS needs to be updated in the following ways:
- Resolving the http vs https in identifiers, and provide same-as references
- Moving from a URI-based identifier to IRI-based identifiers, to simplify encodings
- Clearly specifying how to handle items that contain spaces in the IRIs (current thinking: use underscores-replace-spaces consistently)
- Syncing the endpoint updates with the above items when the data is updated by the editors
- Providing an automated update to the full-text and semantic search environment in the endpoint when the data is updated by the editors.
This is being worked on, and we hope to share news as soon as this is done.
If there are any professionals working in this field that would like to help in this regard, it would of course be very much appreciated.