Issue 418: SPARQL Endpoints
Posted by Béatrice Markhoff on 17/5/2019
Does someone knows if there is somewhere a list of public CIDOC CRM based datasets accessible by programs via a SPARQL Endpoint?
I saw this was already asked by Stephen Stead on 9/10/2014, and some of the listed endpoints are alive:
British Museum : https://collection.britishmuseum.org/resource/sparql
Yale Centre for British Art (YCBA) SPARQL Endpoint : http://collection.britishart.yale.edu/sparql/
OpenArcheo HumaNum :
I'm sure there are many others now: do you think that the CIDOC CRM web site would be a good place to have such list maintained?
Posted by Thomas Francart on 17/5/2019
http://data.doremus.org/sparql provides FRBRoo data on classical music.
Posted by George Bruseker on 17/5/2019
I like the idea of listing endpoints on the site myself. That being said, I would think we would want a bit of metadata around it (obviously) and should make clear that we aren't necessarily endorsing (SIG wouldn't have time or resources to vet deeply). Then the question is would there be a policy about which endpoints get listed? I imagine the provider would have to send their endpoint and metadata to the list and SIG would have to decide whether it makes a minimal grade. Would there be objective criteria or just the decision of the group? If we didn't set some kind of bar, then we would open ourselves to having to place any endpoint on the site. Thoughts?
Posted by Martin on 17/5/2019
I think providers should consult the chapter about compatibility in CRM version 6, and declare what classes and properties they use/ support. That is easy and straight forward. Further, a small text on the kind of material that can be retrieved. The only check would be, if the claim is correct.
We could then think of using some collection-level metadata format.
Posted by George Bruseker on 18/5/2019
I propose to make this an issue for the next sig to take a decision on.