Closed mellybelly closed 6 years ago
I agree that DOIs do not solve the attribution problem. Sometimes people don't even report the name of the ontology, let alone any kind of identifier, URL, paper, or whatever.
Here's a really obnoxious idea. What if we had a tool people could use on their CC-BY ontology that would autopopulate annotation assertion axioms for some generally well-exposed annotation property (e.g. rdfs:label) with some brief verbiage of link that encouraged the user to think more deeply about attribution.
I remember a while back when SWISS-PROT would have a ginormous blob of licensing in every fasta header, that was obnoxious.
This issue was moved to OBOFoundry/OBOFoundry.github.io#592
Recently I was asked to report on who is using our ontology work. Turns out that is very hard. I go to meetings and see people referencing our content, but its not discoverable via Google. It is not in the LOD cloud. How can we do better here? For many open source software projects we can track downloads or have users register. But ontologies are meant to be free and easy to use with no restrictions. Unfortunately this puts us at a disadvantage when justifying their need to funding agencies.
See some ideas on use of DOIs for releases here: https://github.com/oborel/obo-relations/issues/83
This might be a more stable handle for archiving a release with Zenodo, and I like that part, but it is also duplicating the function of the PURLs. We are trying it on for size for the RO. (see other ticket)
I am not sure that a DOI really solves the problem of attribution though. This is more of a licensing issue? Or maybe the new site can track users? but this doesn't solve the problem of many ontologies being in other distribution registries.
@tpoisot and @cmungall may have ideas