> > The examples above do not provide a convincing motivation for such a > development, since we can reduce them to standard elimnators using > well known techniques. > >
I think the Majority function is a good example to convince you. (We can have many similar examples). maj :: Bool -> Bool -> Bool -> Bool maj True True True = True maj True False b = b maj False b True = b maj b True False = b maj False False False = False For this function, there is NOT a way to define it as the form maj x y z = Elim_Bool ....... and satisfy all the five equations (computationally). But if we allow partially defined functions and pattern matching, the Majority function can be defined as it is. Yong ============================== Dr. Yong Luo Computing Laboratory University of Kent