All CODAS processing of UHDAS data now uses:
- numpy for numerical calulations
- matplotlib for graphics
- basemap for map projections and topography
If you use python–and you should–then install IPython. This provides a powerful interactive Python shell and is extremely valuable for beginners and experts alike.
Mercurial is a nice distributed version control system that we are now using extensively. It is written in python. It is compact, easy to install, easy to use, and fast. Much of our software is available using mercurial, served on this web page: http://currents.soest.hawaii.edu/hg .
SCons (http://www.scons.org/) is the new build system that we are using for the core CODAS3 code, in place of makefiles. Like Mercurial, SCons is written in python.
We use Cython to create C from Python.
If you can, just switch to Linux. Ubuntu makes it easy.
Cygwin from Cygnus (now RedHat) (Gnu port) Unix-style tools and environment for Windows, including a bash shell.
MinGW and MSYS are simpler alternatives to Cygwin; we use a version of the MinGW gcc compiler for compiling CODAS3 Windows binaries.
Latex is a powerful markup language that treats math equations particularly well.
Unzip and Zip from InfoZip. These are excellent portable compression and archiving utilities. We use them for our software distribution. You probably already have either unzip or a Windows equivalent; if not, you can download and install them. Linux distributions may not install zip and unzip by default, but most likely include them.
Although we have switched entirely to Python (for data analysis, plotting, ADCP processing, and more) these comments about Matlab are still valid.
Rich Pawlowicz wrote m_map, an excellent, free Matlab mapping toolbox that we use extensively. Although it is not required for ADCP processing, your editing plots will be fastly improved if you have it. Follow the instructions in Section 9 or 10 to install topography. You may have to fiddle with the default renderer (‘zbuffer’ or ‘painters’) to get topography to show up – perhaps different versions of matlab vary in how they treat patches?
We recommend using Postscript format (or Encapsulated Postscript) for saving Matlab plots. Free Postscript interpreters and viewers are readily available; again, Unix-like systems usually have them installed by default, but for Windows you will need to download both Ghostscript and GSView.
