Can Dataverse store Metadata *only* (and no actual datasets)?

[Beth]

 Hi,

I'm looking for a metadata repository to act as an internal card catalog of existing datasets we have in house.  At this moment centralized storage of data is not a requirement for us.

Is there a way to use Dataverse without loading/sharing actual datasets, or is that really a primary requirement?   If so, is there any guidance out there about how to configure, etc?

Thanks for your thoughts,

Beth

[Sherry Lake]

Hi Beth,

UVa's Dataverse repository allows "metadata-only" records. It is reflected in our repository collections policy: https://wwwstatic.lib.virginia.edu/files/2018-03/LibraCollectionsPolicy-20180122.pdf

Items accepted...... Descriptions and links to persistent identifiers (e.g., DOI) for products of research stored in discipline-based...

And is mentioned in Step 7 in our deposit checklist: https://www.library.virginia.edu/libra/datasets/libra-data-deposit-checklist/ 

We also allow uploading of files as well.

Best,

Shery Lake

--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/8906fb60-2f91-4fb1-aff9-853062d33809n%40googlegroups.com.

[Thomas Jouneau]

Hi

This is directly connected to the ongoing discussion in the registrymd WG.

For Sherry : I read your links and I think I have a question : in the case of a metadata-only record you generate a new DOI, isn't it? If the data to which you link your metadata already has a DOI (or another PID) in the external repository, doesn't the same dataset receive two distinct DOIs/PIDs? Or did I misunderstand your process?

Best

Thomas

To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/CADL9p-U9hJTbhza8c65Kd93oswODJEUeY_XrDZsWwE8SdN92SA%40mail.gmail.com.

[Philip Durbin]

Yes, lately we've been using the term "registry" when we mean that the data is somewhere else (not uploaded directly to Dataverse). As Thomas indicated, there have been a few meetings on this topic and I thought I'd include a link to the notes: https://docs.google.com/document/d/1cqFXU8mnGsgbkxJ1p3WYR__TIdkPHKh0uOx8mxZ3iqU/edit?usp=sharing

To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/4bb19dd3-cee2-45a7-8713-3008cc01a3eb%40gmail.com.

--

Philip Durbin
Software Developer for http://dataverse.org
http://www.iq.harvard.edu/people/philip-durbin

[Valentina Pasquale]

Hi Sherry, hi all,

I would also be interested to know what Thomas asked about assigning a new DOI to datasets that should already have one assigned elsewhere. We were also thinking to give the possibility to our users to create metadata-only records that point to another source, but I would like to be sure of the correct way of doing this. Is it ok to generate a new DOI? Where should the original DOI be included in the metadata? Is the "Other ID" field a suitable one?

Thanks for your help!

Best wishes,

Valentina

[Sherry Lake]

Hello  Valentina & all,

You ask a great question that I have been thinking about: why are we "assigning a new DOI to datasets that should already have one assigned elsewhere".

Main answer is because we are using Dataverse software to create metadata-only records, and that with Dataverse software you must assign permanent identifiers, either ALL handles or ALL DOIs. We have no choice but to give metadata-only records DOIs like our other dataset records (those with files). [This is what we want to do and this is how the software works.]

But this does not answer your question should metadata-only records be assigned DOIs? I'm still thinking about the best answer for this.

One thing I am sure of, if the dataset record does not have the file(s), the original DOI should be included in the metadata and "Other ID" is the most appropriate field. In regards to DataCIte metadata, that metadata schema does have a "relatedItem" field. I do not know if the metadata from "Other Id" is sent to DataCite or not, but that might be something that could be added to the DOI minting?

Great conversation!!

Sherry Lake

To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/77e5ab68-9858-46ff-b13e-07048d434abfn%40googlegroups.com.

[Julian Gautier]

Hi everyone,

I agree, if the metadata of datasets published in another repository are also published in a Dataverse repository, the ID from that other repository should be included in the Dataverse repository. And currently using Other ID seems most appropriate to me, too.

The Dataverse software doesn't send to DataCite what's entered in the "Other ID" field. In the OpenAIRE export (another metadata export based on the DataCite standard), "Other ID" is mapped to DataCite's AlternateIdentifier field. The relationship between that AlternateIdentifier field and that RelatedIdentifier field that Sherry mentioned is interesting. The RelatedIdentifier field has mandatory properties to describe the type of relationship. So wouldn't "AlternateIdentifier" be the same as using RelatedIdentifier with the IsIdenticalTo or IsVariantFormOf properties? If so, I wonder if DataCite has considered removing AlternateIdentifier from their metadata standard or why they may not (besides making backwards incompatible changes to the standard...).

Anyway, I hope I haven't strayed too far from your original question, Beth. Adding to what Thomas, Phil and Sherry have written, the Dataverse software has primarily been designed for storing the metadata and files of datasets (and particular datasets that have never been published elsewhere). There are ways now to publish the metadata of datasets published elsewhere, and with the "registries" work the Dataverse community is working on better supporting and standardizing practices for doing that.

And to tie in that hairy metadata question, support and standardizing should eventually include how best to store and export the IDs that other repositories have given those datasets.

Best regards,

Julian

[Philip Durbin]

We're happy to talk about DOIs, Handles, PIDs, etc. here but I wanted to point out that NISO hosts an excellent PID Forum at https://www.pidforum.org and they have topics like "PID Best Practices" and such.

So please feel free to continue discussing here but a thread could also be started there for more on best practices, etc.

Thanks,

Phil

--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/9c279e1e-105b-47f1-94e2-a2065fcfc10an%40googlegroups.com.