[R-pkg-devel] Possible open-source license incompatibilities within R packages

Ilmari Tamminen ||m@r|@t@mm|nen @end|ng |rom |c|oud@com
Mon Sep 22 14:44:48 CEST 2025


Hi, could the moderators remove my previous message. For some reason my email client screwed the response, as you can see:

https://stat.ethz.ch/pipermail/r-package-devel/2025q3/012000.html

I didn't find any address to moderators. Here is the previous message, hopefully better this time:


Sorry about my unclear communication. In the context of solving the assumed license conflicts in the lme4, I meant that you (Ben) could ask only one permission for the minqa, one permission for numderiv, and so on. The maintainers of those packages should know who contributed to the packages, and if there are multiple people, the maintainers could act behalf of those (for giving centralized permissions to you). However, this (and what follows) is only my best layman's understanding currently.


It would indeed nice to have some earlier credible applications of the GPLv2 section 10. or relevant FSF FAQ. For example, to me the expression "parts" does not exclude "all parts". And would "all parts" even used in a usual case, if only a limited number of functions and their arguments are called from an R package. The packages are not usually even distributed along the code of interest (in this case the lme4), why saying "all parts were incorporated" does not sound valid. Even if the lme4 would call all the functions and their arguments from a package, the "incorporated parts" would still be limited to those superficial elements. And this could be agreed with the dependency maintainers: only the function calls are incorporated (the situation now resembling the namespace-API interpretation, but enhanced with the permissions).


The spirit of the FSF in the section 10. is also notable. In the lme4 case, "the parts" are incorporated with the GPL-2 | GPL-3 software, promoting. sharing and reusing.


But these are only my hopefull interpretations. As said, it was surprising to find the section 10. so late, nobody haven't talked about it. Coming searches should be focused on the application of the section 10. I will search more when I have time.


If you feel that getting the lme4 dependency licenses updated is possible, then I think it would be the clearest option from these two.


Regards
Ilmari Tamminen


More information about the R-package-devel mailing list