Guile linear algebra and the GNU Scientific Library

[Mark Galassi <rosalia@cygnus.com>]

    >> > 1) Is anyone interested in this package (if no one is interested,
    >> >    maybe I should not bother).
    >>
    >> Yes, yes, yes, yes!

    Mikael> Yes, yes, yes, yes! from me too :)

A "me too" fest on guile!  I wonder if we could host the list as
guile@aol.com :-)

I am also a strong proponent of that.

I would encourage people looking at numerical methods and very high
level languges (VHLLs) to take a look at GSL (GNU Scientific Library).

There is an earlier alpha snapshot on

  ftp://alpha.gnu.org/gnu/gsl-0.3b.tar.gz

which has a very good random number suite (meant for real physics
research and monte-carlo work) and fft library (a C port of the
classic FFTPACK from netlib), as well as root finding, basic
statistics, error handling, simulated annealing, and so forth.

Development is somewhat active, though the team is small (wanna
join?).  In particular, since the version that's up on alpha.gnu.org,
we have been working on making most of it thread-safe and implementing
special function integration libraries.

A good fraction of GSL is grockked by SWIG directly, and since the gh_
interface has settled down a bit, I'm going to ask David Beazley if he
can update the gh_ support that's in SWIG right now (based on
guile-1.0).

So Chris, I look forward to seeing your linear algebra stuff, and I
wonder if you would be willing to implement it as a C library and then
wrap that for use with Guile (that might be how you are doing it
already).  Once you do that, we could incorporate it into GSL.