Hi,
there is a slowly growing number of (data) repositories who expose metadata encoded in the DataCite metadata schema.
In some cases repositories host copies of resources which are published elsewhere.
E.g. figshare hosts copies of supplementary material of publisher journal articles using the publisher DOI.
So far so good, but I wonder how the identifier and locators should be properly provided in the metadata:
Attached is an example of how this is currently done in figshare DataCite metadata.
The DOI belongs to the publisher and identifies the supplementary material (in this case the video) in the journal article.
Figshare stores a copy of the video and provides the URL locating to the figshare repository.
However the figshare-URL is actually not an alternative identifier of the DOI provided in <identifier>.
Rather there is a specific relation between the publisher version of the resource and the figshare version of the resource. (the same resource but physically stored in two different locations)
So in my opinion the correct way to provide the identifier and relation information in the figshare metadata should be:
I know this is different from the original intended use of the DataCite schema to describe DOI metadata but reflects a typical use case in repositories where the DataCite schema can be used in different contexts.
What are the opinions of the working group experts?
Best regards,
Jochen