mathematics and physics

[gibbs]

I would like to know if there is some project (for ruby) for mathematics
and physics like in python. For example http://www.scipy.org or
http://www.vpython.org/
Thanks.

 

[Axel Etzold]

Datum: Wed, 17 Oct 2007 04:25:07 +0900

Von: gibbs <[email protected]>
An: (e-mail address removed)
Betreff: mathematics and physics

I would like to know if there is some project (for ruby) for mathematics
and physics like in python. For example http://www.scipy.org or
http://www.vpython.org/
Thanks.

Dear Gibbs,

well, there is Ara Howard's Sciruby

ttp://sciruby.codeforpeople.com/sr.cgi/InterestingProjects

Then, there are many bindings. For numerical tasks, I like Ruby-Gsl, i.e. bindings for the Gnu Scientific Library for Ruby.

http://rb-gsl.rubyforge.org/

There are also bindings for the R statistical language,

http://raa.ruby-lang.org/project/rsruby/

I also like this page (partly in Japanese, not all related to Ruby):

http://www.math.kobe-u.ac.jp/~kodama/index-e.html

... and there are many others.

Best regards,

Axel

 

[Jason Bornhoft]

This is a very interesting post for me and I hope that it generates a
lot more interest.

Personally, I would like to see Ruby go as far as LAPACK or BLAS. So
much so, that I think I will end up personally investing time in such a
project (or create my own) in the not-to-distant future.

 

[jzakiya]

This is a very interesting post for me and I hope that it generates a
lot more interest.

Personally, I would like to see Ruby go as far as LAPACK or BLAS. So
much so, that I think I will end up personally investing time in such a
project (or create my own) in the not-to-distant future.

Check out the Maya Project -- astrophysics in Ruby

http://www.artcompsci.org/kali/five-year-plan3.html

 

[jzakiya]

This is a very interesting post for me and I hope that it generates a
lot more interest.

Personally, I would like to see Ruby go as far as LAPACK or BLAS. So
much so, that I think I will end up personally investing time in such a
project (or create my own) in the not-to-distant future.

Check out the Maya Project -- astrophysics in Ruby

http://www.artcompsci.org/kali/five-year-plan3.html

 

[Joel VanderWerf]

Jason Bornhoft wrote:
...

Personally, I would like to see Ruby go as far as LAPACK or BLAS.

...and evolve into a credible replacement for MATLAB/Simulink/Stateflow,
as well. Why not.

 

[Cameron McBride]

This is a very interesting post for me and I hope that it generates a
lot more interest.

Personally, I would like to see Ruby go as far as LAPACK or BLAS. So
much so, that I think I will end up personally investing time in such a
project (or create my own) in the not-to-distant future.

RNum tries to do this: http://rnum.rubyforge.org/
(I haven't used it much personally, hence I say "tries" - it may well succeed).

I've been using NArray and rb-gsl for several years on many scientific projects.

Cameron

 

[Phlip]

Personally, I would like to see Ruby go as far as LAPACK or BLAS.

...and evolve into a credible replacement for MATLAB/Simulink/Stateflow,
as well. Why not.

cue frantic banjo & fiddle music <

Well, maw's out there, switchin' in the kitchen,
And daddy's in the livin' room, grousin' and a-bitchin',
And I'm out here... kicking the gong... for...

----> Euphoria! <----

 

[M. Edward (Ed) Borasky]

This is a very interesting post for me and I hope that it generates a
lot more interest.

Personally, I would like to see Ruby go as far as LAPACK or BLAS. So
much so, that I think I will end up personally investing time in such a
project (or create my own) in the not-to-distant future.

I've been chipping away at the edges of this for a couple of years. From
what I've seen, once the Ruby 1.9 virtual machine/engine is stable,
we'll be able to talk about doing science in Ruby without having to hook
into off-board C libraries. But until that day, your best bet is

a. Learn SWIG (http://www.swig.org)
b. Find a high-performance C library that does what you want to do.
c. Build your interface and enjoy the fruits of SWIG's labors.

It's easy enough to interface Ruby to libraries directly, of course, but
the nice part about SWIG is that with one interface descriptor, you can
drive the library from any reasonable scripting language: Ruby, Python,
Perl, PHP, Lua, Pike, several implementations of Scheme, Java and a few
others I've forgotten.

But Python does have a one- or two-year jump on Ruby in this area, so
we're likely to be playing catch-up. But Python also seems to have a
commanding lead over Perl -- other than the Perl Data Language, which is
mostly for image processing, there isn't much in Perl.

 

[M. Edward (Ed) Borasky]

Jason Bornhoft wrote:
...

...and evolve into a credible replacement for MATLAB/Simulink/Stateflow,
as well. Why not.

Lots of folks would like to "replace" MATLAB. I don't think Ruby and
MATLAB have a heck of a lot in common. And there are lots of open source
math packages, both numeric and symbolic. I personally prefer doing math
in a higher-level language than Ruby, such as R or Axiom.

In any event, I don't think the horse to beat is MATLAB any more, but
Mathematica. MATLAB is really a numeric processing package -- it's
popular in signal and image processing, but there are better tools for
just about everything else these days. Mathematica, on the other hand,
is a really first-rate hunk of code. I've never bought a license for it
-- I don't do enough symbolic math in my day job to justify the cost and
at home, I'd rather play with open source.

 

[Joel VanderWerf]

Lots of folks would like to "replace" MATLAB. I don't think Ruby and
MATLAB have a heck of a lot in common. And there are lots of open source
math packages, both numeric and symbolic. I personally prefer doing math
in a higher-level language than Ruby, such as R or Axiom.

In any event, I don't think the horse to beat is MATLAB any more, but
Mathematica. MATLAB is really a numeric processing package -- it's
popular in signal and image processing, but there are better tools for
just about everything else these days. Mathematica, on the other hand,
is a really first-rate hunk of code. I've never bought a license for it
-- I don't do enough symbolic math in my day job to justify the cost and
at home, I'd rather play with open source.

Unfortunately, a lot of people are doing (and teaching!) numerical
simulation using Simulink/Stateflow. It's not a very elegant language,
IMO. And it doesn't scale well to large, self-reconfiguring networks of
components.