Closed ptsefton closed 1 year ago
I can confirm it. I have fetched the contents from https://www.w3.org/ns/iana/link-relations/relation , but it does not contain a #cite-as definition in the form of triples.
I am not sure where cite-as might be being used, but this term looks like it might be a useful replacement/addition,
https://id.loc.gov/ontologies/bibframe.html#p_preferredCitation
@stain what do you think, can we use preferredCitation from the Library of Congress?
I think our namespace for link relations is wrong or outdated.
https://www.iana.org/assignments/link-relations/link-relations.xhtml is the official IANA registry, which defines cite-as
as of RFC8574 (April 2019) but W3C's (another Standard body) https://www.w3.org/ns/iana/link-relations/relation was last updated 2 years before that in 2017.
(base) stain@xena11:~$ curl -I 'https://www.w3.org/ns/iana/link-relations/relation' | grep -i last-modified
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 32852 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
last-modified: Tue, 24 Oct 2017 21:31:10 GMT
Pinging @hvdsomp on what may be the right namespace.
@dret also have http://webconcepts.info/concepts/link-relation/cite-as which I think is an unofficial concept registry covering http://webconcepts.info/concepts/link-relation/ -- it has a nicer landing page than the outdated W3 Turtle.
I would prefer to stay with cite-as
so we are aligned with FAIR Signposting which we are using with RO-Crate and FAIR Digital Objects (e.g. WorkfowHub)
We know http://schema.org/citation is a bit weak (it could be a citation for a paper about the thing, or a paper that the thing cites, rather than citable identifier for the RO-Crate as a digital object) and http://schema.org/identifier not specific enough unless you inspect the string for https://doi.org/
(which is too narrow on identifier providers).
BTW https://www.researchobject.org/ro-crate/1.1/metadata.html#recommended-identifiers recommends For a Root Data Entity, an identifier which is RECOMMENDED to be a https://doi.org/ URI
-- using cite-as
would make this explicit which is better signposting.
http://www.iana.org/assignments/relation/cite-as redirects correctly (but not to right row), the namespace http://www.iana.org/assignments/relation/
(note no s in http
) was suggested by RFC5988 and updated in RFC8288 A2 suggested for use by Atom format, -- so perhaps we should use the same here.
The core problem here is that there is no official way to express well-defined IANA link relation types (all are strings) as URIs. This issue has come up time and again and, relatively recently, I have lamented about it in the context of work for RFC9264 Link Set, see https://github.com/ietf-wg-httpapi/linkset/issues/45#issuecomment-925865896. In 2016, it was discussed at length in https://github.com/mnot/I-D/issues/140. The pragmatic approach is the one that @stain proposes, i.e. use http://www.iana.org/assignments/relation/
as the URL to which an IANA link relation type (string) listed at https://www.iana.org/assignments/link-relations/link-relations.xhtml is appended. This is, BTW, also the approach used in COAR Notify as can be seen from the JSON-LD context file used by the protocol. In COAR Notify, FAIR Signposting is also an important building block for interoperability; hence their use of cite-as
, which explicitly has the semantics desired by @ptsefton as can be seen from RFC8574.
Should we add this to the section on the root dataset in addition to setting a DOI in the identifier property?
Also do we need a way to add a TEXT statement of how to cite - something like a datacite citation not a link?
This should be addressed by #255, awaiting review.
Describe the bug The default context has a definition of cite-as which does not resolve. https://www.w3.org/ns/iana/link-relations/relation#cite-as
(I want to use this property to put a text citation on a crate so it can say "Cite this crate as ...")
URL https://www.researchobject.org/ro-crate/1.1/context.jsonld