What are the benefits of applicative parsing over monadic parsing?

Question

There seems to be a consensus that you should use Parsec as an applicative rather than a monad. What are the benefits of applicative parsing over monadic parsing?

Answer 1

Monads are strictly a more featureful abstraction than Applicatives. You could write

instance (Monad m) => Applicative m where
  pure  = return
  (<*>) = ap

But there is no way to write

instance (Applicative a) => Monad a where
  return = pure
  (>>=) = ???

So yes, it is essentially a matter of style. I imagine if you use return and ap, then there should be no performance loss over using pure and <*>. Because Applicative is a strictly smaller interface than Monad, this means that <*> can sometimes be more highly optimized than ap. (But with clever GHC rewrite rules, one can often achieve the same optimizations regardless.)

Is monadic parsing out?

Since Monads are a subset of Applicatives, I would conclude that applicative parsing is a subset of monadic parsing.