talking about functions in Haskell

on SO

TLDR:  there's no way you can define in set theory a function that does not return a result. (Actually you can, but then the whole sets axiomatics should be overhauled.)