The people conducting the test also, without the authors' knowledge, got a grad student, gave him eight days to learn Haskell (with no formal training, but the option to ask a more experienced Haskell programmer questions), and then handed him the problem spec and told him to implement it in Haskell. He did it in eight hours with a bit under twice as many LOC as the paper's authors' version. The paper's authors also point out issues with the study, and describe it as showing the possible worth of functional languages, not Haskell specifically.