Re: [Axiom-developer] Axiom and R

[M. Edward (Ed) Borasky]

What do Axiom users want from a numerical package? Given that Axiom was 
once "married" to the NAG library, I'd expect such things as basic 
linear algebra, ordinary and partial differential equations, digital 
signal processing, and statistics. R has all of these, but it is more -- 
it's a complete environment all by itself. And quite a lot of the wisdom 
of R is embedded in contributed packages rather than in the base 
environment.

R can function as a command line interpreter, a shared library (DLL on 
Windows as well), a DCOM server on Windows, a CGI interpreter, and even 
a web application server! There are lots of ways to quickly get some 
kind of interface up, especially if you're willing to do "inelegant" 
things like spawn a process, open up a command line interpreter and pipe 
R language to it.

As an R programmer, I think what *I* would use most would be the ability 
to use Axiom as a high-level language to generate R code, just like the 
"ancient" CAS systems did with FORTRAN code. The R language is a lot 
better than FORTRAN as a target language. And I don't have enough 
experience with Axiom (or any other CAS except Derive) to know what 
sorts of problems I could pose that I couldn't get solved in a timely 
manner in Axiom alone, without a numerical "assistant" like R.

Looking at the other way around -- embedding a CAS in R -- of course, R 
can spawn an Axiom process and pipe code to its standard input. I'm not 
sure I'd take that approach, since R is capable of directly linking to 
C, C++ and Fortran code. I'd be more inclined to take the Ginac symbolic 
calculator library (written in C++) and just do a straighforward and 
largely mechanical integration.

Pierre Doucy wrote:

> Martin,
>
> depending on what you actually want to do, you might just want to make
> them talk to each other through, for instance, a pipe (and some
> translation in between) that might do the trick if what you want to do
> just needs basic types.
> That might also be a good start to understand what are the constraints
> and what needs to be done to actually embed axiom in R (or the other
> way around).
>
> Pierre
>
>
> On 19 Dec 2005 15:58:34 +0100, Martin Rubey <address@hidden> wrote:
>  
>
> > Dear Ed,
> >
> > in fact, a connection between R and Axiom would be wonderful and this was
> > already stated a few times.
> >
> > I also do not really know which way to go, embed Axiom in R or embed R in
> > Axiom...
> >
> > Do you have ideas?
> >
> > Martin
> >
> >
> >
> > _______________________________________________
> > Axiom-developer mailing list
> > address@hidden
> > http://lists.nongnu.org/mailman/listinfo/axiom-developer
> >
> >    
>
>
> --
> Cats are intended to teach us that not everything in nature has a function.
>
>  

--
M. Edward (Ed) Borasky

http://linuxcapacityplanning.com