Top-level vs. local recursion
Compare the following two implementations of map. The first one uses top-level recursion
map :: (a -> b) -> [a] -> [b] map _  =  map f (x:xs) = f x : map f xs
whereas the second one uses local recursion
map :: (a -> b) -> [a] -> [b] map f = let go  =  go (x:xs) = f x : go xs in go
Although the first version is shorter, there are reasons to prefer the second version for stylistic reasons:
- It clearly shows that the 'f' is not "altered" in the recursion.
- You cannot accidentally jump into the wrong loop. I often implement the same function more than once for comparison. The implementations may differ in laziness or performance. Their names are only slightly different and it happened too often that after copy&paste&name-change the recursive call went to the original (and thus wrong) function.
- The local loop is probably also more efficient in execution, because the compiler does not need to move the around. However, if this is actually more efficient then the compiler should do such a transformation itself.f
which is both short and allows deforestation.
- Haskell-Cafe on HLint 1.2