On Mon, 24 Nov 2003, James Amundson wrote:
> On Fri, 2003-11-21 at 09:57, Richard Fateman wrote:
> > R has a larger installed user base than Matlab/Octave?
> > That is pretty surprising to me... are there user
> > statistics somewhere? Is this the Xlisp-stat system
> > underneath it? Can you provide a URL for it?
>
> I don't know about Matlab and Octave combined, but it's pretty clear to
> me that R has a larger user base than Octave alone. See
> www.r-project.org. R is to S/Splus as Octave is to Matlab. (I say that
> as a non-expert.)
I'm basing my estimate partly on the raves I get from statisticians I
meet and partly on the vitality of usenet groups and mailing
lists, but also on a >20-year old estimate we made when I was teaching
applied maths. We wanted to know how people who use the maths they study
in college so we could improve our offerings. Of the people who use maths
in their work, the vast majority use statistical tools (not just academic,
there are many working in medicine and commerce). R has very high
penetration (many people who have S-plus at work use R at home, and those
who use other tools for work but don't have S-plus often have R as well).
You can't say that about Matlab and octave. Since my estimate was done
before computer science took off, I could be wrong if matlab has heavy
penetration there.
I think the X-lisp engine for R was rewritten from scratch. There are
many things you can do in Matlab but not in octave. The few things you
can do in S-plus but not in R are mostly things you shouldn't be doing
anyway (e.g., procedures known to have serious flaws). The main practical
difference is the memory allocation strategy -- with R you have a limit
on memory that is defined at startup time (you can run it in a
multi-user student lab setting with control over memory usage).
--
George N. White III <aa056@chebucto.ns.ca>
Head of St. Margarets Bay, Nova Scotia, Canada