The ocaml runtime is wonderfully simple (http://rwmj.wordpress.com/2009/08/04/ocaml-internals/ http://rwmj.wordpress.com/2009/08/04/ocaml-internals/) and the co

The ocaml runtime is wonderfully simple (http://rwmj.wordpress.com/2009/08/04/ocaml-internals/ http://rwmj.wordpress.com/2009/08/04/ocaml-internals/) and the compiler focuses mostly on good codegen rather than advanced optimisations. In practice, it's common to be able to look at a block of code and predict how fast it will be and how much memory it will allocate. This is not my experience in haskell - the difference in performance when certain optimisations do or don't trigger is huge and is hard to predict just by reading the source code.