Re: Finding source code for stl (was: [S] FORTRAN PROCEDURE)

Prof Brian Ripley (
Thu, 18 Jun 1998 10:18:33 +0100 (BST)

I've changed the subject matter to something more meaningful. Once
again, please do use helpful subject lines. It really helps
when working remotely across slow or flaky connections.

> From: Oliver Klinger <>
> Through an inquiry at MathSoft I just learned that the C- an
> FORTRAN-source-codes are not available because they may not leave MathSoft.
> Therefore there are no ways to list internal source code.

That is true for S-PLUS code, but I believe you can buy a source
code licence for S, if you try hard enough. However, stl is in
SHOME/splus/.Functions, so is presumably an S-PLUS addition to or
modification of S. Sometimes inspired guesswork will reveal the
source of the code: for example the code inside kmeans is the
Applied Statistics algorithm 136, and that inside ppreg is Friedman's
SMART code.

> But: There is a program call "R" which is a freely distributable
> S+-reprogramming!!! I'm not sure but I think there's free source code
> available for R. MathSoft told me that the programming interface is the same
> for R and S+, so there might be similar functions you could view AND edit.
> I'm going to try it myself and if I get important results I'll post again.
> So - stay tuned

That's sort of true. However, R and S are not the same and internally
are quite different, so examining the source code for R (which is all
public) will not necessarily tell you how S works. More seriously, R
implements a fairly small subset of S-PLUS, with minimal time-series
functionality and definitely not including stl. R has its own mailing
lists, and discussion on R is better placed there than here.

Brian D. Ripley,        
Professor of Applied Statistics,
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

----------------------------------------------------------------------- This message was distributed by To unsubscribe send e-mail to with the BODY of the message: unsubscribe s-news