[R] licensing of R packages

[Carlos Ungil]

Duncan Murdoch-2 wrote:
> 
> The way to lose a GPL lawsuit is to incorporate GPL'd code into your own 
> project, and then not follow the GPL when you redistribute.  There's 
> evidence of that.
> 
> But I've never heard of anyone linking to but not distributing GPL'd 
> code and being sued for it, let alone losing lawsuits over it.  That's 
> evidence enough for me that it is a safe thing to do.
> 

The LGPL covers the first point as well as the GPL does (I think), while
explicitily allowing dynamic linking (so the second point is clearly not a
problem). The FSF encourages using the GPL (and not the LGPL) precisely to
make libraries available only to GPL projects. It's not surprising therefore
that the GPL license scares people off.

The safest option is to take the GPL at face value and accept the FSF
interpretation, but depending on the jurisdiction, the details of the
situation, and the level of risk-aversion, people might decide to do
otherwise. The concept of "derivative work" is not really well defined, see
for example http://www.rosenlaw.com/html/GPL.PDF

To give another example related to R, the FSF foundation view of the world
is that RPy, because it links dynamically to libR.so (or R.dll, etc), has to
be distributed with a GPL (-compatible?) license. And the same restriction
applies in turn to a python program using RPy (again, according to the FSF;
because RPy and the "derivative work" are dynamically linked by the python
interpreter).

However, not everyone shares that view:

http://mail.python.org/pipermail/python-list/2005-January/304974.html
> On the basis of these clauses, the legal advice to us was that merely 
> including "import rpy" and making calls to RPy-wrapped R functions does 
> not invoke the provisions of the GPL because these calls only relate to 
> run-time linking, which is not covered by the GPL. However, combining 
> GPLed source code or static linking would invoke the GPL provisions. 
> [....]
> IANAL, and the above constitutes mangled paraphrasing of carefully 
> worded formal legal advice, the scope of which was restricted to 
> Australian law. However, the sections of the GPL quoted above are pretty 
> unambiguous.
> The other, informal advice, was to ignore the FAQs and other opinions on 
> the FSF web site regarding intepretation of the GPL - it's only the 
> license text which counts.

Cheers,

Carlos
-- 
View this message in context: http://www.nabble.com/licensing-of-R-packages-tp20497391p20509444.html
Sent from the R help mailing list archive at Nabble.com.