Issue 18096

doi link renders awkward and is not working

Reporter: mdoering
Type: Feedback
Summary: doi link renders awkward and is not working
Resolution: Fixed
Status: Closed
Created: 2015-12-17 19:58:39.497
Updated: 2017-10-05 14:33:50.606
Resolved: 2017-10-05 14:33:50.584
Description: There are 2 DOIs for this dataset. One from GBIF linking back to this dataset page and one from zenodo which links to the paper.

The Zenodo one is the main one and renders nicely - just as expected.
The GBIF one hides unter alt ids, looks bad and might not be needed at all, is it?]]>

Author: hoefft
Comment: I would say no, but i guess it depends on wether they both serve a purpose. Do we need both? Do we want users to use the original or the new one. Or is there a use case for both? If there is a need for both, then it should be clear which to use when.
Created: 2015-12-21 10:30:53.86
Updated: 2015-12-21 10:30:53.86

Comment: Once we minted a GBIF DOI it basically needs to remain forever. As a dataset might be known under several identifiers (not only DOIs) we list them as alternate identifiers. There clearly should only be one primary, current DOI, but we should show the others too if they did exist once.
Created: 2015-12-21 12:47:46.919
Updated: 2015-12-21 12:47:46.919

Author: hoefft
Comment: But does it makes sense to add a gbif doi in the first place if the original dataset has one already. When would this be useful. Only if we want to reference/use our own doi elsewhere i assume? Is the publisher DOI for the actual data or could it simply be the description of the dataset. In that case i guess it makes sense that we keep a gbif doi for reference. If not shouldn't we just use the existing one for everything?
Created: 2015-12-21 15:06:30.704
Updated: 2015-12-21 15:06:30.704

Comment: We only create a GBIF DOI if the dataset did not have one when it was first registered. When an updated metadata document comes in with an externally created DOI we switch to use that, but keep the previous GBIF one around
Created: 2015-12-21 15:56:51.187
Updated: 2015-12-21 15:56:51.187

Created: 2016-06-03 10:05:22.908
Updated: 2016-06-03 10:05:22.908
Actually, due to POR-3116 we are assigning DOIs to datasets during dataset registration/creation, when they already have a DOI assigned to them in their EML. At least for datasets published via the IPT, a workaround can be done, see

As Markus says, we need to show all DOIs for a dataset and make sure they render properly.

To make it clearer which DOI was assigned by GBIF, perhaps we can better distinguish it from the publisher-assigned DOI somehow? [~hoefft] []?

Author: hoefft
Comment: both DOI shows and is formatted. the publisher DOI takes precedence based on the API I see
Created: 2017-10-05 14:33:50.604
Updated: 2017-10-05 14:33:50.604