Jean-Baptiste G.

But Reducer already extends LeftReducer, there is no notion of fast or tagged at this point (in the method with the Reducer constraint): only at call site it is decided...

So, let's recap: `Reducer` is a type class introduced to improve algorithms efficiency. `LeftReducer` and `RightReducer` could allow to better document the efficiency conditions of the said algorithms. This is...

Also added by name evaluation and Delay type classe. If we want to generalize this then for Foldable, instead of ```scala def foldRight[A, B](fa: F[A], z: => B)(f: (A, =>...

Yes, this a known issue (cf. comments in thread https://github.com/scalaz/scalaz/pull/1450#issuecomment-381231202), but it is good to have a dedicated github issue to track it. Personally I'd be in favor of removing...

@mielientiev what was the workaround that you applied for #1756 to solve the issue? In any case in seems wise that modules should not have any `val`s, ie. either use...

Given #1602, we could also have the following support: ```scala implicit def reducingCofoldable1[A, F[_]](implicit R: Reducer[A, F[A]]): Cofoldable1[F, A] implicit def reducingCofoldable[A, F[_]](implicit R: Reducer[A, F[A]], M: Monoid[F[A]]): Cofoldable[F, A]...

this is definitely useful! For the record, do we want to be able to unfold more complex data structure, like `Tree`s? In which case we will need [something more evolved](https://hackage.haskell.org/package/unfoldable),...

> unfoldr is dual to foldr. by the way, should the default `Cofoldable`/`Unfoldable` instance for `IList` do an `unfoldr` or an `unfoldl`? I believe the later is what is currently...

@vmarquez yes, it should work for `Tree`, did not dig too much either before commenting sorry. Now I understand that the main difference with the haskell package, is that it...

@vmarquez, I think the same laws should also work for the more general `F, A` case (instead of `F[_], A`)