Mike Meyer
So the question is - is there any advantage or harm in this way of looking at values?
One harm is that you are mixing two concepts, which Haskell specifically tries to keep separate: functions and procedures. A function maps a value to another value. A "zeroadic function" hence is not a function. It is indeed a value and the thunk concept is just a technical detail, about which you shouldn't bother. You can write proper Haskell programs without ever knowing, what a thunk is or how lazy evaluation is /implemented/. Now in most strict languages there are no real functions, and procedures are just named "functions" for some reason. In my opinion the "zeroadic function" view is going to hurt you much more than it is going to help you. Brent makes a good point that it may help you, when coming from a strict language, but on the other hand understanding lazy evaluation, non-strictness and forcing will make Haskell much more natural to you. With zeroadic "functions" you may end up not writing natural Haskell programs, but just trying to fit your strict view into the non-strict, lazily evaluated world. Greets, Ertugrul -- nightmare = unsafePerformIO (getWrongWife >>= sex) http://ertes.de/