Simpler enough to use and reason about that tools in such languages have been mathematically verified a lot by relatively small teams. The imperative, popular alternatives are usually PhD projects for people attempting such things. Speaks volumes about the intrinsic complexity of reasoning about such things.