Is there a reason Exception and Maybe don't implement Bifunctor?

[mattly]

Perhaps I'm misunderstanding Monadic laws or the like, but I've hit a situation where it would be useful to bimap over an Exception or Maybe monad without necessarily knowing which type it is, and it seems only Either implements the Bifunctor protocol. Is this for Category Theory reasons or simply an omission?

[yurrriq]

I wasn't able to find any good resources online quickly, but as I understand it, a Bifunctor instance for Exception or Maybe doesn't make sense, because unlike Either or Pair, there's no "left" value to map over.

Nothing represents the absence of a value and Exception represents an error, thus not a desirable value. Therefore it doesn't make sense to want to map over them.

[yurrriq]

If you look at Data.Bifunctor:

Bifunctor Either
Bifunctor (,)
Bifunctor Const
Bifunctor ((,,) x1)
Bifunctor ((,,,) x1 x2)
Bifunctor ((,,,,) x1 x2 x3)
Bifunctor ((,,,,,) x1 x2 x3 x4)
Bifunctor ((,,,,,,) x1 x2 x3 x4 x5)

We've got instances for cats.data.Pair and cats.monad.Either, but maybe it'd be cool to do clojure.lang.PersistentVector or some other seqs too.

[niwinz]

I agree with @yurrriq