I didn't write most of it so I've been cleaning it up as I go, but it's tedious because it's impossible to check completely without a lot of code path analysis. For example, class A has a method called "fred()" and class B extends class A and also has a method called "fred()". You can remove fred() from class B and not get any errors from Eclipse unless some piece of code directly references B.fred() (and in a correctly designed OO environment that shouldn't happen!).
That means the only way to detect an issue is to run the code and ensure that fred() is called. With an application as large as MT -- and which only has a smattering of documentation here and there, and nothing to cover the overall architecture -- that's going to be very error-prone.
