Support some datacite.yml in addition/instead of CITATION.cff

[yarikoptic]

Your idea

ATM there is an idea to relay more of information for citations to CITATION.cff and not strive to make dataset_description.json to contain all needed information (e.g. affiliations of the authors etc).

BUT, as far as I see, CITATION.cff development stalled, while it is still lacking some rudimentary needed features such as supporting multiple affiliations for authors (citation-file-format/citation-file-format#268) and unclear how to reference funders (citation-file-format/citation-file-format#491).

While that is happening, there is https://schema.datacite.org/ with a working group, continuous progress, and used widely by anyone minting DOIs via datacite (e.g. we do for DANDI and IIRC OpenNeuro also uses datacite).

A well aged example of a datacite record from @mih for a BIDS dataset could be found at
https://github.com/psychoinformatics-de/studyforrest-data-phase2/blob/master/datacite.yml . It lacks affiliations but otherwise a good example .

Ideally there should be a tool to convert from datacite to citation.cff format, but ATM I didn't google up any. may be someone could add some support into https://github.com/con/tributors or alike.

[Remi-Gau]

FYI bids2cite was my attempt to help 'converting' info from dataset_description to citation.cff OR datacite.yml...

https://bids2cite.readthedocs.io/en/latest/usage_notes.html#command-line-arguments

[effigies]

It's not ideal to have three ways of doing things, but I agree that CITATION.cff seems to be stagnating and BIDS is constrained by backwards compatibility to support CITATION.cff.

IMO if datacite.yml has the features needed and now, it's worth supporting, and I would be inclined to adopt it for OpenNeuro-generated metadata instead of CITATION.cff, which is still on our roadmap.

cc @nellh for your thoughts.

[effigies]

Update: OpenNeuro would prefer to work with datacite.yml than CITATION.cff at this point, but we would also prefer to support BIDS-only datasets.

Will bring up with maintainers this week and possibly try to add to steering agenda for next week.

[effigies]

A result of today's maintainer's call was general agreement that supporting multiple citation file formats is reasonable to improve the FAIR-sharing of BIDS datasets that are deposited in public databases. We would prefer that the different formats be mutually exclusive within a dataset, as duplicated metadata would need to be validated or risk becoming contradictory.