Setting up OS X as a Scientific Programming Environment (From MacResearch)

From MacResearch (http://www.macresearch.org/setting_up_os_x_as_a_scientific_programming_environment):

I came across this great little article in setting up a scientific programming environment on OS X written by Dr. Eric P. Salathé Jr., a researcher in the Department of Atmospheric Sciences at the University of Washington. There are a number of sites with simple "essential application" lists for OS X, but this page is focused specifically on scientific programming, and instead of presenting a mundane list of applications, Dr. Salathé goes into detail on why each application should be installed and the scenarios in which they are used. I also like the fact that he's included several gems from the open-sources world that some Mac users may be unaware of. Although Dr. Salathé personal interests lie in the Atmospheric Sciences domain, he lays out a capable scientific programming environment that is suitable for any number of scientific disciplines. Topics covered in the page include:

• Installing fortran and C compilers
• Dealing with binary data
• Installing Optimized math libraries

You can view the complete article on Dr. Salathé's Website

Dr. Salathé's site looks incredibly useful. Dealing with Fortran on OS X is a battle. While I don't use Fortran for my own work, it is a dependency of R. I have a custom built R and Rpy installation for development of the Chartshare project. Every time I upgrade either component, there is at least a couple hours involved in fighting with GNU Fortran.

Comments

[Patrick]

The two big things I feel he left out (I'm writing a Wiki entry for our first Wiki entry on this same subject) is selecting the install option to install XWin during upgrade or installation and immediately grabbing the most current stable Developer's Toolkit before you do anything else.

Thanks to your advice I've found that if you do that then whatever else you want to do later will usually fall into place (until you start compiling R!). I've got this box (banjo) tricked out, including fully funtioning F77/F90 and GCC, the binary data handling mods that he writes about along with R, Mathematica, Matlab, Aqua emacs and TexShop.

I like it now but I'll probably be cussing the next time GCC point releases.