Should metadata be automatically created for DOIs?

[mbjones]

From email thread from Steve Kraffmiller and Greg Janée:

With that problem out of the way, I ran into the next problem, which is that the test code attempts to create a DOI identifier without supplying the required metadata. (This software was apparently written before metadata requirements were put in place some time ago.)

The bottom line is that this client code works, and is probably a good starting point if you're trying to develop a Java solution, but the code is a little dated and needs some updating. You might consider contacting the author of the code.

At this point I don't see a bug here, but I could be convinced otherwise. The EZID library provides a method to submit the metadata along with the create request, so the user of the library is able to submit the required metadata along with the request for the DOI by passing in a hash array of the metadata to be submitted. We don't automatically do this, nor do we require it, because the other schemes like ARKs don't have the same requirements, so we want to maintain the flexibility for users to create identifiers without metadata while supporting the submission of required metadata as well. Is there really anything that needs to be changed here?

[mbjones]

Greg Janée replies:

I think the only issue here is that the user was trying to do a "mvn package", which at the end tries to create a test DOI, which fails, which would flummox any new user. If the test DOI creation supplied some metadata, then the whole package installation would succeed.

Thanks, this does seem to be a bug, so I am retagging this issue. Fixing the tests should be straightforward.