Regarding the utilisation of the GPL2 section 10 for solving the GPL2-only vs. GPL3 conflicts, it is important to remember the GPL3 virality. If only the GPL licenses are being read then it looks like that the conflicts exist, and the two most certain solutions are:
1) Update the upstream licenses 2) Not use the GPL2 packages downstream Then, however, the FSF's GPL FAQ takes the more liberal stance, allowing for example the command-line communication within software aggregates. It doesn't sound a valid independency, if one code entity cannot call another one and then fails executing its purpose because of it, but this kind of application without the GPL3 virality seem to be widely accepted despite the language in the GPL licenses. Furthermore, to my understanding the GPL licenses do not clearly take into account whether the dependencies are conveyed with the code of interest or not, making the situation even more blurry. Thus, the situation is still unfortunately unclear. It seems that the language in the GPL licenses is the strictest, but the popular and widely accepted interpretations are more liberal way or another, divided into the two schools (the R community's way I would like to call as the namespace-API interpretation). Regards Ilmari Tamminen ______________________________________________ [email protected] mailing list https://stat.ethz.ch/mailman/listinfo/r-package-devel
