Copyright information in MIT/BSD-Licenses and mixed licenses

[Iglezakis, Dorothea]

Dear Dataverse community,

we are actually preparing the custom license information of our datasets for the upgrade to version 5.12. We want to migrate as many of the custom terms to standard licenses, including software licenses like MIT, BSD, (A/L)GPL and Apache 2.0.

But how can we handle the (individual) copyright information in MIT or BSD licenses?

And does anyone plan to include mixed licenses for data and code? In our existing datasets we have the cases MIT (Code) + CC-BY (Data), GPL (Code) + CC-BY (Data) and Apache 2.0 (Code) + CC-BY (Data) that would be nice to standardise also.

Any ideas or experiences from any of you? 

I'm also happy to continue the discussion in issue #8512, if preferred

Thanks a lot and best wishes,

Doro (thea Iglezakis)

__

Dr. Dorothea Iglezakis
FoKUS - Kompetenzzentrum für Forschungsdaten
70174 Stuttgart

Email: dorothea....@ub.uni-stuttgart.de

[Dieuwertje Bloemen]

Hi Doro,

At KU Leuven, for mixed licenses for code and data, we advise our researchers to split the dataset in two. One with the code and a code license and one with the data and a data license. These two datasets then refer to each other via the related materials field. We also advise them to register the same related publication if there is one and to have the title be the same with one having "(Replication)Data for:" and the other having "Code for:" in front of it.

I'm personally not a proponent of mixing licenses, as the url of the license can't really be mixed.

Kind regards,

Dieuwertje (KU Leuven RDR)

[Philipp Conzett]

Hi all,

The level of licensing has been discussed on several occasions in the Dataverse community. Right now, I could only find this discussion thread in the Google group.

From previous discussions, I recall that several installations see the need for being able to add standard licenses at file level. In the case different standard licenses are applied within a dataset, the question arises what license the dataset as a whole would need to have. Some argue that standard licenses or customized terms of use only should be applied at file level, whereas the license for dataset as a whole only would cover the metadata about the dataset, for which the recommended license is a CC0; see also the GitHub issue Metadata licensing #6888.

I suggest discussing file-level licensing / standard license support in its own GitHub issue, but cross-link it with Standardize standard license configuration #8512.

Best, Philipp

[Sebastian Karcher]

Hi all,

Yes, file-level licensing would be useful to us.

In the meantime, we have mainly two combinations of licenses -- CC-BY-SA with our standard, non-controlled-access agreement and CC-BY-SA with our controlled access agreement, and we've created meta-license pages for both of those to be able to use DV's standard license feature. So for example if you look at a regular project on QDR like https://data.qdr.syr.edu/dataset.xhtml?persistentId=doi:10.5064/F6IJHYLR and click on the "QDR Standard Access" link, it takes you to: https://qdr.syr.edu/policies/qdr-standard-access-conditions, which in turn provides a plain language summary as well as links to the two applicable licenses. Depending on how many different licenses you manage, this will of course become impossible, but as long as you have only two licenses per projects and only 2-3 per type, this should leave you with a manageable set of combinations. The biggest downside, of course, is the lack of machine-readability: e.g. CC-BY or CC0 data aren't going to be recognized as such if you slap a custom meta license on them because you also want to account for Apache-licensed code. That's fine for our purposes (the data themselves are under our custom licenses anyway), but probably not for more typical set ups with fully open data.

All the best,

Sebastian

Sebastianh

--
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/67cf7b96-e70d-4f6b-b95d-4bdb8224a61bn%40googlegroups.com.

--

Sebastian Karcher, PhD
www.sebastiankarcher.com