> I also think Haskell places an unreasonable burden of thinking about the types vs. the algorithm an domain.
When you are encoding your logic in types the distinction between types and "the algorithm and domain" is very blurred if not the exact same thing. I wouldn't want to be programming in any other language if I knew very little about the domain.
But "encoding your logic in types" is often non-trivial and very different from thinking about the algorithm itself. Of course, most algorithms require dependent types (at the very least!) to be fully encoded, so you can't really encode them in Haskell...
