To me the value of Haskell comes from its being an extreme point in the language design space: that of a pure, lazy, strictly-typed language. I think that Haskell provides a very nice syntax and GHC a very nice implementation for exploring this extreme point. The primary benefit of this exploration is that it enriches our understanding as a community regarding language design and program design. I think that is a worthwhile enough contribution.

I think that Haskell's strangeness harms it in two ways. One it puts off people because the superficial strangeness takes some getting used to. More dangerously it makes a small fraction of Haskell programmers dangerously arrogant. They feel that just because they are now familiar with this strange jargon they have become superior to the users of other languages. Depending on temperament this turns them into snobs or evangelists.

I don't think there is any concrete evidence of Haskell's superiority as a practical tool. Programs like pandoc, darcs or xmonad are great but there is no proof (and as far as I know neither have their authors claimed) that they were easier to develop than comparable programs in imperative languages (TeX, git?).

Much better is the attitude of Haskell leaders like Simon Peyton Jones who present Haskell as a ongoing research project with its set of successes and unsolved problems: http://research.microsoft.com/en-us/um/people/simonpj/papers... http://research.microsoft.com/en-us/um/people/simonpj/papers...