Closed Thibauth closed 3 years ago
So if I understand how FOIA works correctly (I am very likely wrong here), the raw data itself is public domain. Our source code for processing it, as well as the output data files we create via linkage etc can be protected under any license we choose.
I think we should use the CC BY NC license, or at least CC BY , with the additional provisio that the form of attribution is an actual citation to the dataset paper (for both code and our output data).
Thanks for your input. However, it is generally considered a bad practice to use a CC license for code (even CC discourages it here). So we should probably use a different license for code, I don't really have a preference, but I think the MIT license would be fine.
Good point -- MIT license for code sounds fine to me. Do you have any preference for the license of the output cleaned data?
@Thibauth thoughts on c4c48a6 ?
A few comments:
src/
folderIn any case, having a small section in the README, as you have done, is a great way to make clear what is covered by what.
See b945719
I agree about SA and about splitting the files.
I think we should keep the software license in the main folder to indicate that it covers all of the software (not just stuff found in the src
folder). E.g. the Makefile
is also covered under MIT.
@Thibauth if this all looks good you can close this issue!
This is perfect, thanks!
We need to add licensing information to this repository (also, this seems to be required in the NeurIPS call for paper). We might need a different license for the data and for the code, so this is a bit tricky. Any opinion/suggestion is welcome.
Some useful information and guidelines here.