Re: [Rd] Creating RPMs for Packages
This may get delayed as I am away from my usual mail system. Since (thanks to Dirk and Charles) there is the cran2deb project giving us most R packages in debian package format, could the alien package help? See http://www.linuxdocs.org/HOWTOs/RPM-for-Unix-HOWTO-8.html John Nash __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
[Rd] Use of cfortran.h?
I'm planning to try to use Mike Powell's new BOBYQA optimization routine to see if it is a good candidate to eventually replace Nelder-Mead as the optim() default. (Don't panic -- this won't happen quickly.) Mike writes f77 code. After some chasing about, I've realized that because optimization codes need to call the objective function written in R, they need to use a language that can do this. C can, Fortran cannot. There is, however, a package cfortran that allows for mixed coding in C and Fortran. This might not be the ultimate solution -- e.g., it may not be Windows usable. However, it could be a useful tool to try out some programs that are already available in Fortran. My query is to ask if anyone has experience they can share in using cfortran.h. I'd also welcome (off-list may be best) input on directions for improving or at least tidying optim() and many related capabilities in R. John Nash __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] sources of code; was Generate random numbers in Fortran
Ben Bolker gives some reasons why Numerical Recipes may be problematic as a starting point for R codes. CUP did a masterful job of marketing, but the license is restrictive as the links he gives points out. In some tests, I've also noted that some of the algorithms are less than stellar e.g, convergence tests in one or two optimization routines. Should we have a wiki item to help people find material? My own "Compact Numerical Methods: linear algebra and function minimisation" codes were first published 30 years ago this month. They are the Pascal library on Netlib. I have some Fortran codes I could post, and BASIC versions too. In fact, the function minimisation codes that are actually more advanced than the routines Brian Ripley used in optim() have already been made freely available (http://www.nashinfo.com/nlpe.htm). The codes still run right away in DOSBOX along with GWBASIC.EXE, but I believe the more useful aspect is providing methods and ideas. Gnu Scientific Library has been mentioned. It has some strengths but a number of "holes". There are some other notable collections. The Decision Tree for Optimization (Hans Mittelman) is a helpful link, though I am not certain all the resources are unencumbered. Would an annotated list of such openly usable resources be helpful? Perhaps if a couple of folk contact me off-list we can try a wiki item and see if it "works". JN __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
