I didn't know hoogle, thanks for sharing. For statically typed languages that make sense and especially so for functional languages like Haskell. Test-driven search is much more suitable for dynamically typed languages where you exploit duck-typing to your favour, e.g. you don't really care if the parse_bool function in the example takes a str or a bytes type as input and even more importantly you don't care about the type of the returned value as long as it evaluates to True and False.