We need more bindings to scietific libraries in Bigloo or to any Scheme in general

[Förster vom Silberwald]

My wish for the new year: more bindings to scientific libraries.

Actually we have:

- DISLIN binding (Förster vom Silberwald: wholly completed)

- gsl (Will, PhD student from MIT is working in his spare time on a gsl
binding to Bigloo)

- here I would appreciate: a binding to R language (only the core
functionality) or to any statistics package. gsl has some basic
statistics functions but not as many as R.

The other day I was forced to use R. It has a lot of usefull functions
but nobody can tell me that its buil-in programming language has any
virtue. Surely, they write that R is based on some Scheme but they do
not tell why they crippled R and made them without paranetheses.

I once thought adapting X-Lisp to Bigloo would be valuable.

Regards,
Siegfried Gonzi
PS: There exists actually two bindings to R in Python. As I see it:
they used some tricks to circumvent to have to deal with R objects
because realizing types in Python are not that easy.

[synth...@uol.com.br]

Förster vom Silberwald wrote:
> My wish for the new year: more bindings to scientific libraries.

I wonder whatever happened to a plt library for simulation?
Cheers,

Henry

[Jens Axel Søgaard]

You are probably thinking of the impressice PLT Scheme
Science Collection.

<http://planet.plt-scheme.org/#science.plt>

There is a draft of the documentation at:

<http://www.ccs.neu.edu/home/matthias/Tmp/science.pdf>

--
Jens Axel Søgaard

[ihaka]

I'm one of the original R "perpertrators". I have to agree that it is
rubbish, but it does demonstrate what can be done when a bunch of people
who really understand statistical methodology start to work together. The
syntax was bait to get other statisticians to join in and to leverage
existing code (S-Plus). It did the job.

At the moment I'm casting about for a real scheme to use as a base for
statistical work. My shopping list consists of
cross-platform portability (including gui), reasonable efficiency, good
FFI, unicode support, preemptive threads, Clos-like object system. Given
a reasonable subset of this list, I'll be happy to work at bringing as
much functionality from R to scheme as I can. The trick will be to get
other statisticians to chip in.