On Fri, Nov 26, 1999 at 09:16:11AM +0000, Markus Neteler wrote:
> ludcmp routine - finds inverse of matrix A
> lubksb routine - multiplies inverse of matrix A (from lucmp) by matrix b
> and returns matrix x (weights) in "b". =20
>
> I recommend removing these commands from the grass-beta source code
> until some rewrites these basic functions or the copyright situtation
> is clarified. I think the algorithms can be
> found in any good algebra book. Somebody just have to do it.
from my thesis work I can give you a routine for matrix inversion.
Well I haven't volenteered yet to redo it.
Mainly, because I am not familiar with these GRASS modules.
(And time is lacking, as you know. 
)
I have written a SVD (singular value decomposition) routine and
other routines. My other matrix operations are based on
"MESCHACH"
Homepage: ftp://ftpmaths.anu.edu.au/pub/meschach/meschach.html
Seems to be Free Software.
ftp://ftpmaths.anu.edu.au/pub/meschach/README
| The source code is available to be perused, used and passed on
| without cost, while ensuring that the quality of the software is not
| compromised. The software is copyrighted; however, the copyright
| agreement follows in the footsteps of the Free Software Foundation in
| preventing abuse that occurs with totally public domain software.
A lot of other stuff is at: NETLIB
It differs from package to package.
http://www.netlib.no/netlib/master/index.html
-> fftpack
for: Fast Fourier Transform of periodic and other symmetric
sequences
Mostly in Fortran> :-/
But free software at least the C translation:
/* These Fourier routines were originally based on the Fourier
routines of the same names from the NETLIB bihar and fftpack
fortran libraries developed by Paul N. Swarztrauber at the National
Center for Atmospheric Research in Boulder, CO USA. They have been
reimplemented in C and optimized in a few ways for OggSquish. */
/* As the original fortran libraries are public domain, the C Fourier
routines in this file are hereby released to the public domain as
well. The C routines here produce output exactly equivalent to the
original fortran routines. Of particular interest are the facts
that (like the original fortran), these routines can work on
arbitrary length vectors that need not be powers of two in
length. */
-> itpack
for: Iterative Linear System Solvers
Jacobi method, SOR, SSOR with conjugate gradient acceleration
or with Chebyshev (semi-iteration - SI) acceleration.
This one is not explicit about the licensing. 
It probably is, because most researchers published software and
algorithms excatly for the purpose that other can build on it.
(Would defeat the scientific purpose otherwise, wouldn't it? )
The SAL is a good entry point to search for Free Software and numerics.
http://sal.kachinatech.com/B/index.shtml
Bernhard
--
Research Assistant, Geog Dept UM-Milwaukee, USA. (www.uwm.edu/~bernhard)
Free Software Projects and Consulting (intevation.net)
Association for a Free Informational Infrastructure (ffii.org)