Skip to content

Regarding {KernSmooth} - Can a package on CRAN have non GPL copyrights?

4 messages · Tal Galili, Joshua Wiley, Murray Stokely +1 more

Original
Joshua Wiley
# 
AFAIK CRAN does not explicitly require use of a GPL license (though
obviously the license needs to allow at least free access to the
packages and source).
On Sat, Feb 12, 2011 at 11:57 AM, Tal Galili <tal.galili at gmail.com> wrote:
Well, I think that is just a warning that it was built under a
different version of R than you are currently running it under.

If memory serves, the current KernSmooth is a port to R by Brian
Ripley; that message suggests the original KernSmooth was copyrighted
by M. P. Wand from 1997 until 2009 (probably the original author in
S?).

Cheers,

Josh

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Murray Stokely
    

    

      
      #
    

  


  

      
On Sat, Feb 12, 2011 at 11:57 AM, Tal Galili <tal.galili at gmail.com> wrote:

            

      
      
      
    

        
There are many licenses used on CRAN, such as BSD, MIT, GPL, LGPL, and
public domain.  There are even some that aren't really open source at
all such as 'akima' and 'SparseM'.  These packages make additional
restrictions such as that a license is only granted for academic /
non-commercial use.  It's particularly annoying that many other
packages without such restrictions depend on these packages, making it
pretty difficult to ensure license compliance for companies using R.

The fact that the code is Copyright by that author is somewhat
orthogonal to the license he chooses to distribute it under.  See the
LICENSE file.  In this case, it's much more open than GPL -- it is
completely unrestricted / public domain.  After reading the license
for the package you are interested in, you may like to read how those
licenses are interpreted by e.g.

   http://www.opensource.org/licenses/index.html

             - Murray

  
  


  


      

    

    

      
      

        


  

        

      Brian Ripley
    

    

      
      #
    

  


  

      
On Sat, 12 Feb 2011, Joshua Wiley wrote:

        

            

      
      
      
    

        
GPL and the other licences are about distribution.  CRAN requires a 
licence to allow it to distribute packages -- in some cases (e.g. 
mclust) that requires a CRAN exception.

Copyright is an almost entirely orthogonal issue.  In many countries 
all intellectual creation is copyright -- this message is, R itself 
is, my packages are (including my contributions to KernSmooth) -- and 
that does not need to stated.  Matt Wand has chosen to assert his 
copyright in the startup message.  A closely related concept you will 
see in British books is that 'X has asserted his moral right to be 
identified as the author of this work'.

With the freedoms that Open Source licences give goes a moral 
responsibility to give due credit, something we see **far** too little 
of here.