> The imperative approach really is simpler to wrap one's head around

I've devised a way of programming imperatively in Haskell but in a way that maintains all the benefits of the type system ("make invalid behaviours unrepresentable") but I don't think that moved the needle on Haskell's popularity...