I 2nd that...
Probably the shortest path in terms of development time by Michael and the FMS team would be to make a FMS State Server which would dump out positions,velocities and orientations of the model to a specified port or something.
Once the server is in place, this would allow students (like Daniel and I) to write software for our specific application which would listen to the FMS State Server, use whatever mathmatical package we have at our disposal to calculate desired control inputs, and then output control inputs to any one of the appropriate ports from which FMS can recieve controller inputs.
I believe that this scheme will work, will be fairly easy to implement by the FMS programming team and is general enough to allow for multiple applications since some students may want to use MatLAB while others may want to use their own custom software.
PS. I just started using FMS for learning how to fly (or more importantly how not crash) helis, so thankyou!