using GMP's algorithms in other projects

[Torbjorn Granlund]

Paul Leyland <paul at leyland.vispa.com> writes:

  On Fri, 2008-08-29 at 13:09 +0200, Torbjorn Granlund wrote:
  > Vincent Diepeveen <diep at xs4all.nl> writes:
  > 
  >   Is it allowed to use the code from GMP for that?
  >      Answer = yes, but if you do that then you must also publish your
  >      code.
  >   
  > This is a misunderstanding.  You do not need to publish derivate
  > works.  But if you do distribute them, you have to publish them under
  > the (L)GPL.
  
  Is that *really* the case?
  
  Suppose I write some code which makes non-trivial calls to GMP and
  distribute it only as source under a license whereby I explicitly forbid
  redistribution of my own code in any form other than source.  The
  imposition of that condition means my license can not be {L,GPL}
  compatible.  On the other hand, recipients of my code can easily make
  derivative works which they can't then distribute, despite what the
  {L,}GPL would like them to be able to do.
  
I don't follow, but then again, I tend to be very little interested in
hypothetical lawyer's problems.

My main point is that you can play with (L)GPL source as much as you
want, including making the most perfect version of GMP, and leave that
on your hard drive till kingdom come.

-- 
Torbjörn