How Many RORs Do We Need?

Adopting organizational identifiers (RORs) in existing metadata may not be as hard as we think for DataCite members that only have a few affiliations in their metadata. However, even in these cases, some challenges remain. See https://www.tedhabermann.com/blog/2019/11/10/how-many-rors-do-we-need for more details.

1 Like

Would it be worth ROR utilising other existing domain specific practices to be able to link/complement existing infrastructure and practices? i.e. EDMO (https://www.seadatanet.org/Metadata/EDMO-Organisations)

James,

The ROR database contributed by Digital Science includes identifiers from several other lists for many RORs. There are two types of external_ids in the data: all and preferred. I am not sure what the details of this classification are, but all is generally a list of ids and preferred is either null or a single id. The counts of these ids in the current data dump are shown below. Of course, the preponderance of GRIDs reflects the source of the data more than a preference for these ids.

Source All Preferred Source All
GRID 96793 96793 CNRS 830
ISNI 38004 33 UKPRN 170
Wikidata 26935 456 HESA 168
OrgRef 14870 512 UCAS 153
FundRef 10051 995

The EDMO looks like a great source of organizations in the marine domain. As far as I can tell, EDMO does not include a unique identifier for these organizations other than the EDMO record id. This id could be used as a “compact identifier”, e.g. edmo:4889 for the Cooperative Institute for Research in Environmental Sciences (CIRES) or the EDMO organizations could be mapped to RORs (https://ror.org/00bdqav06 in this case) to provide a connection from EDMO to the broader scientific community.

This is an interesting role for ROR - connecting domain registries to the broader community. I think this would be a worthwhile contribution to both communities.

Ted

Ted,

Sorry for the delay in responding. I agree it would be beneficial for both communities.

Just to let you know EDMO does have a unique ID per organisation i.e. BODC and a SPARQL endpoint and all the information can be queried with machine readable information encoded in RDF.

Thanks
James

James,

I agree that EDMO has an identifier for each organization but I was perhaps a bit unclear in my wording. Sorry about that. I should have said: The unique identifier in EDMO is the record ID.

This id is a good example of a “compact identifier”, i.e. an identifier valid in a well-known namespace.

Ted