Feedback Requested: Add DOIs to older studies that have an HDL as a persistent identifier

[Eleni Castro]

Hi Dataverse Community,

We would like your feedback on the following:

About over a year ago now, at the Harvard Dataverse, we switched to assigning DataCite EZID DOIs as the persistent identifier for any new studies created from then on in. However we have thousands of older studies which have Handles as their persistent ID. We recently got a question on twitter from a Dataverse user asking if we could add a DOI to these studies as well (see conversation here: https://twitter.com/KyleCranmer/status/517083033401180160). 

Should we retroactively add DOIs to older studies that have Handles? What is the added value to creating DOIs for these older studies? Regardless, please note, that we will continue maintaining any Handles we minted for these studies. 

Thanks so much for any feedback you might be able to offer!

Best regards,

Eleni

-- 

Eleni Castro

Research Coordinator, Data Acquisition and Archiving, Data Science

IQSS, Harvard University

http://www.iq.harvard.edu/people/eleni-castro 

[August Muench]

I am glad this came up. I think that updating past handles to DOIs has the same motivation/value as having switched from handles to datacite DOIs in the first place. I know I would prefer it if all the Harvard/CfA subnetwork dataverse were updated to DOIs. 

Are there any metadata mapping issues that will crop up though for the old dataverses/datasets to convert to EZID dois?  

Perhaps it could be on demand and by dataverse. For example, the Murray archive doesn't care about DOIs given that they are one of the larger Dataverses it might be of no value for them to do so. 

 - gus

[Jon Crabtree]

Eleni

My opinion would be yes we should do it for preservation purposes

The risk of adding second identifiers is less than the risk that Handles could fail to resolve in the future if DOIs continue to gain market

The complication would be studies that currently have handles but under new appraisal strategies would not qualify for DOIs

For now the best I feel would be to give DOIs to all existing handles

Going forward deciding when to mint doi's  for new deposits based on other factors will be critical and I know that is on the schedule for 4.0

I  also interested on other opinions since I will face this soon

Jon Crabtree

Odum Institute

Sent from my iPhone

--
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 post to this group, send email to dataverse...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/6e03beea-4a16-41df-b1f7-7deda86e4f16%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Stephen Marks]

I more or less agree with Jon. I'm not a big fan of the way the handle system works, since it relies on the local handle server to resolve the resource, and I've found the server to be a bit finicky, in addition to just being something you will *always* have to maintain. I much prefer the DOI model, where a central registrar assigns and resolves the DOIs.

I'm grappling with how we are going to do this here when we eventually transition to version 4. My thoughts are, we definitely want to deprecate handles to the extent possible (so, stop assigning them, and assign DOIs to datasets that already have handles). We can't do much about handles that are already out in the world in print and etc., but we can try to arrest their spread. We'll continue to maintain our local handle server for now, because it's the responsible thing to do, but I'm not happy about it. =)

To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/2408BEB2-FA9C-42D1-8C7B-9F747010CFD1%40gmail.com.

[Kyle Cranmer]

Hello,

Thank you for taking up this issue. I'm not sure that I have much more to add. My main motivation is that most of the integrations that I see for data (tracking citations, alt metrics, bibliographic tools, etc.) seem to support DOIs but often not handles. I'm glad to see the discussion so far is positive.

Kyle Cranmer

NYU