ddubin65
commented
4 years ago The difficulty with your example, it seems to me, is that appearing first in the author list is not a reliable justification for the gndo:firstAuthor property you have assigned. One answer to this problem has been to adopt the Contributor Roles Ontology. See for example, these:
https://wkauthorservices.editage.com/resources/author-resource-review/2018/may-2018.html
https://onlinelibrary.wiley.com/doi/full/10.1002/leap.1210
I spoke a bit at RDA P9 and P10 about the problem of representing CRediT roles in RDF. The PROV community is fine with shared roles as instances. You and I might serve in the very same project administration roles at different times, or even on different projects. But communities that base their ontologies on BFO (VIVO-ISF, for example) see roles differently. Your project administration role could never be one and the same role as mine, precisely because it is yours. In the BFO view, our respective roles would be instances of the same class. See, for example, the discussion here: https://wiki.lyrasis.org/display/VTDA/Representing+relationships+and+roles.
So whether a CRediT role like 'Writing -- original draft' is a class or an instance depends on who you need to share the descriptions with. My P10 slides are online here: https://github.com/RDAProvPatWG/PlenaryMaterials/tree/master/Plenary10.
I would really be interested in some feedback from the Attribution IG about this issue regarding linking TDWG standards documents to their contributors. At the present, I'm linking people via
dcterms:contributoras you can see here. But this does not capture the nature of the relationships.What I've done so far is to take the relationships recorded in this table and generate this RDF/Turtle which is available to anyone who discovers that dump but won't actually show up if the document URIs are dereferenced (see this example).
Anyway, I'm open to suggestion on the best practices for making these links in a manner that will indicate the contributor's role.