Hi Thanks to everyone that have helped, the runtimes for the haskell programs have decreased significantly, presently they look like this compared to O'Caml: Benchmark haskell ocaml drop3 5.786 3.151 five11 8.657 7.692 huffman 7.134 18.593 uudecode 6.042 2.657 uuencode 7.775 2.824 The huffman benchmark should not really be that slow, but it seems that O'Caml programs that build large lists tend to be slow. The versions that I have used in this measurement is available at: http://bitbenches.infogami.com/ Many thanks to the people that have responded, and if you have more comments on the programs, particularily uudecode and uuencode (which I have mostly written myself), feel free to suggest further improvments. Note also that uudecode and uuencode for both O'Caml and Haskell breaks the rules slightly by creating an array for each row that is created. To end up with a list of rows that is than turned into one array. Is there a library fuction that can do this? I wrote a function myself, but it is not very pretty. Per