This is a great idea. I can think of three key points that are missing from the document.
1) This is an opportunity to finally develop a common sequence file format (xml I would assume). I have said in the past that people who develop new file formats should be shot (why .meg Sudir??), but if everyone can agree to one it will save MANY wasted hours.
2) We need an interface that works on a Mac, a PC and on any operating system (Java? or something browser based) and can funciton on a modest desktop. This should be a priority. Think about the developing world too.
3) most of the problem with the programs arose because they were part of someone's dissertation and not written by professional programmers, but instead by beginning molecular phylogeneticists. They are not used by people who necessarily understand how they work nor the broader problems that they might have been better designed for. Keep this every user is a beginner attitude, and make it a priority to Keep it Simple!!!!! The interfaces must be self explanatory and educational, i.e. an undergrad should be able to learn to do phylogenetics just by using it by reading the menus and instructions.