User:Andreas Plank/CETAF Stable Identifier Handbook

How can I implement CETAF Stable Identifiers for my collection?

The CETAF Stable Identifiers can be implemented in three levels. They are described in detail in herbal.rbge.info’s documentation. Following conditions have to be met to reach the corresponding implementation levels:

(Level 0: identifier choosen)

Level 1

• you assigned a stable URI to each object of your collection, which will be never changed and preferably follows the best practices for stable URIs
• there exists a human-readable representation (web-page) for each of your collection objects
• a user trying to access a collection object by typing the stable URI of it into a web-browser will be redirected to the human-readable representation (web-page) of the object (you can test this by using the CETAF URI Tester)

Level 2

• you reached level 1
• there exists a machine-readable RDF metadata record for each of your collection objects
• a machine trying to access a collection object via its identifier with 'application/rdf+xml' header will be redirected to the objects machine-readable RDF metadata record (you can test this by using the CETAF URI Tester)

Level 3

• you reached level 2
• the machine-readable RDF metadata record of each of your collection objects encodes application specific data (e.g. is compliant to the CETAF Specimen Preview Profile—CSPP)

HTTP vs. HTTPS versions of CETAF URIs

As far as the Semantic web is concerned http://xyz and https://xyz are different things because they are different URIs. The recommendation for new implementations should be just to use HTTPS. If you have only HTTP or HTTPS versions, or want to change it you should take notice of the following:

HTTP	HTTPS
You have issued only HTTP versions of CETAF URIs and want to keep it that way have nothing to add technically, just have the usual 303 HTTP redirect to RDF or HTML resources in place	You have issued only HTTPS versions of CETAF URIs don’t need to resolve then HTTP if you have never issued any, because they aren’t out there to be resolved.
Want to change HTTP to HTTPS
You have issued HTTP versions of CETAF URIs but want to change to HTTPS have to keep resolving with a 303 redirect to HTTPS of the RDF or HTML resources. The RDF should contain an owl:sameAs assertion linking the HTTP and HTTPS versions of the URI, therefore only minor configure stuff for providers and transparent for users. could change to telling people to cite HTTPS rather than HTTP for your specimens but it shouldn’t matter too much as these things are linked together. The recommendation would be to cite as HTTPS if you have it implemented as at some point in the future a client may refuse to trust even a redirect from an HTTP URI (which is a bit paranoid but may happen).

Publishing CETAF IDs to GBIF

If your institution is using CETAF IDs and yout want them (and potential Specimen RDF) to be included into the CETAF Specimen Catalogue, they need to be used as GUIDs in the specimen data fed to GBIF. As described in CETAF Specimen Catalogue, the GBIF Index is used to discover CETAF IDs.

• If DarwinCore is used, the IDs must be mapped to occurrence ID.
• For ABCD, the concept UnitGUID should be used.

How can I discover specimens with CETAF IDs and corresponding LOD?

You can discover specimens of institutions of the Stable Identifiers Implementers Group by using the CETAF Specimen Catalogue maintained at the BGBM, which offers a web service for getting a list of valid CETAF IDs. For implementers of level 2, who provide RDF representations of their specimens, a cache triple store with a SPARQL access point will be available soon.