Compatible Licenses for Different PALS Projects

[DavidSagan]

Since this is a community effort, the license needs to be switched from the LBL one. Any suggestions as to what is best?

[EZoni]

I had asked this question on our internal Slack about two months ago,

Another thing I don't understand in general is, if the project is a collaborative project between LBNL, Cornell, etc., shouldn't there be a collaborative copyright (instead of a LBNL-only copyright)?

and I believe @ax3l responded something along these lines:

We intentionally asked IPO for release permissions before other people contribute to pals-python to simplify the process of initial licensing. It's no problem now that we have the license that more people contribute (under the LBNL/UC license). Generally we have a process in IPO to coordinate with partners if they had already contributed before we do initial licensing, but it takes longer.

In this case, there were, and there are so far, only LBNL contributions to pals-python. Meaning, that only contributors with LBNL affiliation had, and have so far, committed code to the main branch. So we requested a LBNL/UC license. My understanding is that under this license others can contribute further.

Personally, I would have no objection revisiting this, if needed for some reason, but I wouldn't know where this leads a priori.

[DavidSagan]

OK so if pals-python license is to remain as is what should the license be for pals-cpp and pals-julia?

[ax3l]

Thanks for the summary, @EZoni.
Yes, this is essentially just a 3-clause-BSD license with government use clause if forks are published publicly. Any future contributions are the same widely compatible license (3-clause-BSD), anyone can contribute or fork or use a private copy for their needs.

For other projects not lead by LBNL, I suggest any MIT / 3-clause-BSD / Apache2 license will do for widest compatibility 👍

In practice, the people that originate it and do the initial release will need to go through their IP process / IPO to get approval. Differs of course widely per institution.

The original copyright holder is then the one creating the project (or more likely: their employer) and the license gives everyone the right to use and modify/contribute.
One can periodically update the holder if significant contributions/changes were made, e.g., to ack growing institutional engagement formally. That is more of a bureaucratic act that takes a few months of IP people taking to each other. And/or we can also credit scientifically in the README/docs.