[Help-gsl] MT19937 licensing in GSL RNG and beyond

[Paul C. Leopardi]

Hi all,
I'm planning to release a modified version of the Fortran 95 MT19937 code of 
Jose Rui Faustino de Sousa, under the GPL, as part of a random number 
generator interface library. In preparation for this, I noticed that in 2001, 
the authors of the C version of MT (Makoto Matsumoto and Takuji Nishimura) 
changed the license from GPL to a BSD-style license: 

http://www.math.sci.hiroshima-u.ac.jp/~m-mat/MT/MT2002/elicense.html
"Until 2001/4/6, MT had been distributed under GNU Public License, but after 
2001/4/6, we decided to let MT be used for any purpose, including commercial 
use. 2002-versions mt19937ar.c, mt19937ar-cok.c are considered to be usable 
freely."

See also the opening comment in:
http://www.math.sci.hiroshima-u.ac.jp/~m-mat/MT/MT2002/CODES/mt19937ar.c

In contrast, the author of the MT code in GSL RNG (Brian Gough) released the 
code in 1998 under the LGPL, and then changed the license from LGPL to GPL.
See eg. gsl-1.8/rng/mt.c

1. How was it possible to release the GSL RNG version of MT under the LGPL in 
1998, when the license for the Japanese version was not made more liberal 
until 2001? Did Brian Gough need to ask Makoto Matsumoto and Takuji Nishimura 
for special permission needed to do this? Or was licensing under LGPL 
possible because the GSL RNG version was a separate and independent 
implementation of the algorithm, and thus not subject to the copyright of the 
Japanese code?

2. The original version of Jose Rui Faustino de Sousa's Fortran 95 source code 
is no longer available (it contains a bug). I've asked and obtained his 
permission to publish a fixed version under the GPL. He also has a fixed and 
enhanced version available at 
http://www.math.sci.hiroshima-u.ac.jp/~m-mat/MT/VERSIONS/FORTRAN/mt95.f90
Would my release of the fixed version of Jose's original Fortran 95 under the 
GPL be going against the spirit of the more liberal licensing of the Japanese 
version of MT? If so, could I license this one file under LGPL and the rest 
of the pseudorandom number generator interface library under the GPL?

Your thoughts?
-- 
Paul Leopardi, School of Physics, University of Sydney