I attempt to write and maintain some bioinformatics software.
My language of choice is C++, because it runs fast, but I am currently doing much development in Java because that part is quick, with a view to translating to C++ the compute-intensive bits.
My favourite program, and the cause of most of my headaches, is TreeMap, which is a Macintosh (Classic) application desiged to uncover the relationships among ecologically linked organisms.
In another life I created Spectrum, which can be used to show phylogenetic information in a data set, without requiring you to estimate an evolutionary tree. I think this has been largely superceded by other software such as SplitsTree.
CoSpec is another application that models cophylogenetic evolution very simply: I used it to demonstrate the large range of possible coevolutionary patterns that can arise.
Joshua Ho has written GNLab, which is a program for large-scale analysis of Gene Regulatory Networks. It is now compiled for and moderately tested on Windows and Mac OS X platforms.
Markov invariants are useful objects for phylogenetic inference (see http://arxiv.org/abs/0711.3503).
Filo is a Java program designed to read in simple #NEXUS-like files and generate artificial molecular sequence data according to some very general constraints. It was supported by ARC grant DP0770991.
Ben Drinkwater has implemented a faster implementation of the original cophylogeny node mapping algorithm originally proposed by Libeskind-Hadas and Charleston, 2009. The source code can be downloaded as an eclipse project here. This implementation offers a significant running time improvement from the initial implementation of O(n7) to a new worst case running time of O(n3)).
Dr Michael Woodhams has made extensive modifications to the PAL library at the heart of BEAST and we include the updated code here for the public under the same LGPL licence. This work was supported by ARC grant DP0770991.