Wow, this complex.... Thank you. I will try that.
No, don't! There is an easier way.
(...) This is the point at which you are forced to move away from OO thinking.
This seems to be worth clearing a little bit. I usually read people saying that this kind of misundestanding is a confusion with object oriented background, when actually I believe it is not. Magicloud's assumption are all correct, in the Haskell way of thinking. Classes do behave the way he thinks it does. The problem was only that you can't (without existential quantification) build lists of elements with diferent types for each element -- as you also can't do in languages like C++. The way to achieve that in C++ is to use a list of pointers, while in Haskell you can use existential quantification. (Confusion with OO is less of a problem than people use to think it is. I've already been told that something I wanted was OO specific, and I later realized what I needed is available in Haskell and is called associated type, as in type families.) Best, Maurício