The one you linked is close to how you'd write this in idiomatic Haskell. The paper is describing a particular technique, and using FizzBuzz as an example problem, not suggesting that the technique is a good way to solve FizzBuzz.