R-beta: Problems porting nlme to R

[Douglas Bates]

Gerry Brush <gerry at wright.iba.ox.ac.uk> writes:

> I have been trying to port the extension 'nlme' ("Classes and methods to fit  
> and display linear and nonlinear mixed effects models.")to R found on the  
> Statlib S site.  I am having some difficulty.  The c code requires at least  
> S.h and nonlin.h.  R has S.h but not nonlin.h.  I'm only a recent convert to  
> R, and have not used S nor SPlus.

You are a brave person to try to port all that to R :-)

It would be extremely difficult to port nlme 2.1 to R.  We are working
on a version 3.0 of nlme and at least the lme part may eventually be
used under R.  It will still be difficult to port the entire nlme
library, though. 

The nonlinear optimization approach in S/S-PLUS is very different from
that available in R which is why nonlin.h is not available in R.  The
nonlinear optimization in nlme is closely tied to the evaluation model
in S/S-PLUS.  The evaluation model in R is fundamentally different.

Trying to take the nonlinear optimization approach of S/S-PLUS and
incorporate it into R would be difficult.  It is probably better to
redesign the nonlinear optimization from scratch and take advantage of
the evaluation model and the scoping of variables in R.  At least for
me, though, this will not be a project that will happen soon.

-- 
Douglas Bates                            bates at stat.wisc.edu
Statistics Department                    608/262-2598
University of Wisconsin - Madison        http://www.stat.wisc.edu/~bates/
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._