Adam Esterline a twisted journey through my mind

One Last Fibonacci Sequence

I have become obsessed with the Fibonacci sequence in Haskell. Let’s look at one last implementation I found on the “Let vs. Where” page and I swear I won’t talk about Fibonacci any more.

fibFast3 :: Int -> Int
fibFast3 = (map fib' [0 ..] !!)
             fib' 0 = 0
             fib' 1 = 1
             fib' n = fibFast3 (n - 1) + fibFast3 (n - 2)

I like this implementation the best out of all the implementations so far. I don’t think this implementation is the fastest, but I think it is a good mix between fast and readable. Let’s break down the first line of this function.

(map fib' [0 ..] !!)

  • map is the traditional map function. map takes a function and a “list” and applies the function to the list. Keep in mind map is lazy and once a value has been caculated it remembers the result.
  • fib' is our mapping function. fib' is defined just like the simple/slow Fibonacci definition from my first post about Haskell Fibonacci numbers. This simple definition is what I like the most. It doesn’t require any knowledge of Haskell to understand.
  • [0 ..] is our lazy infinite number sequence.
  • !! is our “get” function for lists. [1, 0, 4] !! 0 => 1

To make the whole thing work we need a little currying. That is why we have () around the whole expression. (map fib' [0 ..] !!) returns a function that when given a single integer will “get” that fibonacci number in our fibonacci sequence map fib' [0..].

What do you think? Which one do you think reads the best?