"Apfelmus, Heinrich"
[...] but this is very different from using a particular monad like the state monad and hoping that using it somehow gives an insight into the problem domain.
You're right, mostly. However, there are a lot of problems, where you cannot provide any useful abstraction, or the abstraction would destroy the convenience and clarity of expressing the problem as something as simple as a stateful computation. The 'insight' into a problem often comes from expressing its solution, not the problem itself. Please consider that I'm talking about real-world applications, so my problems are things like internal database servers. Of course, there may be better ways to model such a thing, but a 'StateT (Map a b) IO' computation is the way to go for someone, who wants to get the job done rather than doing research, and in fact I think this is a very beautiful and elegant approach exploiting Haskell's (or at least GHC's) great RTS features. Greets, Ertugrul. -- nightmare = unsafePerformIO (getWrongWife >>= sex) http://blog.ertes.de/