Hi,

There is a version of my R embedding code at:

http://web.kuicr.kyoto-u.ac.jp/~alexg/files

This is very very rough and early. Think pre-pre-alpha. It is a
conversion of the core part of RSPerl allowing you to call R functions
from ruby scripts and pass simple data between them. Simple data means:
integers, floats and strings and arrays and hashes thereof. So you can
write:

require 'rsruby'
r=RSRuby.new #initialize embedded R interpreter
samples = 100
normal = r.rnorm(samples) #Calls R function 'rnorm'
puts normal #a Ruby array of floats converted from R vector
r.plot({ 'x' => normal,          #You can call with named args
         'xlab' => 'Hello Ruby',
	 'ylab' => 'Ruby ga suki desu'})
sleep(4) #Ooooh - pretty!

At this level it doesn't provide anything that Gordon/Ed's suggestion
of using an external R interpreter couldn't do. But perhaps with a bit
more work it might.

Anyway, if you're interested feel free to download and test. I've tried
it on a linux box and OSX, I have no idea how to build it under Windows
(or even if you could!). Documentation is essentially non-existent at
this stage, though there's a README which tells you how to build - you
just need to tell extconf.rb where your libR sits.

You can find my email address for bug reports/contributions in the
DESCRIPTION file.

Cheers,

AlexG