I filed my thesis proposal for my PhD on February 15th 2010.
The real part of the proposal (the last chapter) is willingly short and generic. I believe in an agile approach to planning: you can’t know upfront everything you are going to do, so planning every tiny detail in advance is a useless waste of time. While you do research you get a deeper understanding of the subject (as in programming), and so you get new ideas of trash old ones.
A nice thing to do would be to transform the state-of-the-art chapter in a comparative survey. To do this I need to experimentally evaluate most of the softwares I review. It is for sure not a quick thing. And I also have to find a good benchmark.
Here the pdf of my thesis proposal