Comment on license section

[FlorianPargent]

I have just read the license section and learned a lot! Unfortunately, I think it is quite intimidating and I fear it is still not clear for the average reader with the average project what to do (it also is not fully clear to myself, as I have not spent enough time thinking about licenses myself). I try to list some of my thoughts:

• For our intended readership, I assume an average project contains 1) some data the author is allowed to use (because it is their own or published under an appropriate license) or links to a data repository in the documentation if the data cannot be shared, 2) some scripting code to analyse data using open source software like R and its packages, 3) some literate programming code like Quarto code, 4) some outputs like PDFs of a manuscript or figures, 5) some metadata like README files for the project.
• I am not sure the average reader (including me) knows how to license the project after reading the tutorial. Some things which might still be unclear (or could be made more clear) might be:
  • When you talk about "Software" in the tutorial, does this include 1) every piece of R code I wrote to analyse my data, 2) every piece of Quarto etc. code I wrote in order to make my writing reproducible?
  • Is there a difference between some code I wrote that I can expect to really be useful to others (e.g. I wrote a new function for some specific analysis step that others might use exactly like I wrote it), and code that I wrote to perform my analysis and produce my reproducible manuscript but I would not expect to be useful to others?
  • Assume that I as the average author want to publish a research compendium which is something like reproducible supplementary files to my research article that will hopefully soon be published in a journal. 1) I want to include a permissive license to signal others that they can do with my project whatever they want. In theory, I do not need an attribution license because scientific norms require my colleagues to cite me anyways if they used my materials in their own research (although most researchers might still want attribution to be included in the license). 2) I want to make sure that I do not accidentally license anything under a license I am not allowed to. I expect that readers of my work assume good faith on my part with respect to this point because I am a researcher and not a software engineer and I have not developed "anything" new except from the analyses and interpretation of results I have outlined in my research manuscript. Can you give some advice on shortcutting your flowcharts in this case (I suppose not, but maybe there is). I admit that in my own projects, I have in the past always just added a CC-BY license to my project without further explanation, although I know that this is not really how this licensing stuff works but I always hoped that at least no real harm will come from that?
• Could you include in the wrap-up of the licensing section something that prevents readers to be too frightened by the license choice? Readers' fear to make mistakes in licensing should not prevent them from publishing reproducible materials for their research!

[FlorianPargent]

Let me rephrase one of my thoughts from above:

1. What exactly is the reason why I cannot just select CC0 for the whole project without specifying additional details?
2. If 1 is not possible, what exactly is the reason why I cannot instead write something like "everything in this project which is not already licensed otherwise is licensed under CC0"?

[fkohrt]

This is a very good point. I added a TL;DR via 3f2f2e8.

I will leave this issue open because (1) I am interested whether this adequately addresses your second bullet point, and (2) your third bullet point still needs to be addressed.

[fkohrt]

Regarding (2), I added the following statement as an example to the TL;DR via 45c8a91:

Except where noted otherwise, all files in this project are made available under CC0 1.0 or (at your option) under the terms of the Apache Software License 2.0.

@FlorianPargent do you think all your points are now addressed?

The problem with "just using CC0" is that it explicitly reserves patents (i.e., does not grant patent rights) and, for example, Fedora removed CC0 from its allowed licenses for code. So I think it's better to always dual-license by default.

[FlorianPargent]

I have done some reading and now better understand the problem of licensing code under CC0. I think you should explicitly mention this in your section on dual-licensing. Perhaps something like: "In theory it could be dangerous to use code published under CC0, because there could still be a patent on the code and the patent holder could try to sue you for royalties if you used the CC0 licensed code in your own software project."

[fkohrt]

I factored out the part on dual-licensing in a separate tip via e71addf and added the relevant bit about patents in a footnote.

I would like to focus on licensing own material, as using and adapting material by others is such a large separate topic (although of course, the tutorial already touches on it).

[fkohrt]

Thinking about this again, your example sentence...

In theory it could be dangerous to use code published under CC0, because there could still be a patent on the code and the patent holder could try to sue you for royalties if you used the CC0 licensed code in your own software project.

...belongs to the section where it is explained when people can use existing material. Currently it says that either free or open is sufficient, but I could add a note to exercise caution if a license is not both (as is the case with CC0).

[fkohrt]

If I find the time, I will add a note about using work by others (see #22). Apart from that, I think this issue can be closed.