I find it a bit curious to contrast Haskell and ML -- maybe it's because I did a course on programming languages twice, first with Standard ML as (one of) the "functional" languages, and again when Haskell had taken over that role (the course first (extensively) used Sethi's "Programming Languages").

I've always seen Haskell as a kind of ML off-shot, but perhaps the similarity is more in the syntax than in the substance?