search

src-d / datasets

source{d} datasets ("big code") for source code analysis and machine learning on source code
Other
323 stars 82 forks source link

Change license to ODBL #65

Open campoy opened 6 years ago

campoy commented 6 years ago

As per @eiso's comment in https://github.com/src-d/guide/issues/204#issuecomment-386540245 we should change the current license (Apache v2) to ODBL.

vmarkovtsev commented 6 years ago

@campoy License of what exactly? Tools or index? Index is currently CC-BY-*

campoy commented 6 years ago

The data of Public Git Archive itself, which is downloaded with pga. I guess pga might be Apache v2 but it should be said somewhere that the contents are not.

smola commented 6 years ago

Yes, tools need to preserve Apache 2. ODBL is only applicable to datasets (both index and siva files, but not the contents of siva file themselves such as commits and blobs, which preserve their own license and that needs to be noted too).

vmarkovtsev commented 6 years ago

The index file is Creative Commons Attribution-NonCommercialShareAlike 4.0 International as stated in the paper. Let's not change it until we present the paper on MSR.

smola commented 6 years ago

@vmarkovtsev AFAIK ODbl is less restrictive than Creative Commons Attribution-NonCommercialShareAlike 4.0 International, but if you want to be sure, just keep double licensing on the index.

eiso commented 6 years ago

Final to dual license but we should change the licenses before releasing

smola commented 6 years ago

Also keep in mind that you can add an additional license (as in double licensing) after releasing. But you cannot withdraw these licenses after releasing (or you can, but it probably has no legal effect).

@eiso If I understand correctly, you meant that we should double-license the existing version, and use just ODBL for next releases, right?

eiso commented 6 years ago

Correct, do the dual license on the index file (ODBL & Creative Commons Attribution-NonCommercialShareAlike 4.0 International), so that Vadim doesn't have to be concerned about the MSR paper, and future version ODBL only.

campoy commented 6 years ago

This was not done yet and it's pretty important. If we agree the license is ODbL it should be mentioned somewhere in the README before we release this further.

vmarkovtsev commented 6 years ago

I think all have agreed.

campoy commented 6 years ago

We might have all agreed, but I don't see any reference to any licenses other than Apache v2. Did I miss them?

smola commented 6 years ago

@campoy As far as I see, it was done only for Identifiers dataset, so this is just waiting for @campoy or @vmarkovtsev to do it ;)