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

[Dirk Eddelbuettel]

Hi Henrik,

People have been wrong on the internet before, and they continue to do so.
Just because someone made a claim (which you and I considered false and
without underlying merit), and even when action resulted, does not create
"case law". That was simple a mistake being made. [1]

The underlying issue is that nobody has shown that a Depends (or these days
Imports) implies that the 'importing package' (here your `afex` package)
inherits anything from the 'imported package' (here Ben's `lme4`). CRAN has
over twenty-two thousand packages. There simply is no cross-pollination of
licensing on package being used in this standard way. [2] What matters is
that when a package gets onto CRAN is that its license is a suitable,
understood and accepted license that is part of the master list CRAN
keeps. All that is in the manuals.

I will now try to extract myself now from this thread. It really is a
flashback to earlier days in that it creates a lot of heat without much light.

Cheers, Dirk


[1] And having been part of Debian for thirty years I am with you that
sometimes it is easier to just make a change to a package or file and move
on. That by now means settles the underlying issue. There are simply too many
hotheads to debate.
[2] As said before, it is difference when a package builds, compiles, links,
with an external library. That is the example of virality that
matters. Different case though.

-- 
dirk.eddelbuettel.com | @eddelbuettel | edd using debian.org