add EitherNev alias and syntax

[eugkhp]

I'm using Vector a lot in my 2.13.7 codebase since it's one of the best collections right now and we either need to have those aliases in the business code or convert to List/Chanin in order to use cats' syntax. After asking if it exists somewhere @armanbilge said we can add it to cats.

[armanbilge]

CI doesn't seem to be running 🤔

[eugkhp]

Looks like CI finished 🙌

[eugkhp]

Added both add IorNev and ValidatedNev

[johnynek]

actually... I wonder if we are moving in the wrong direction here.

Like, what if instead we did:

def toValidatedA[F[_]: Applicative](e: Either[A, B]): Validated[F[A], B] =
  e match {
    case Right(b) => Validated.valid(b)
    case Left(a) => Validated.invalid(Applicative[F].pure(a))
  }

then add a low priority typeclass instance for Applicative[Validated[F[A], *]] given a SemigroupK[F].

Then all these instances would be the same.

[eugkhp]

Sure, writing that boilerplate was kinda annoying, especially trying to find all existing ones and make them consistent. I can prepare another MR or add to the existing one. What do you prefer?

[armanbilge]

@key-eugene are you still interested to work on Oscar's proposal? Seems like an interesting idea, I guess a separate PR would make more sense.

[satorg]

Yeah, I like that idea too. Seems similar to the discussion we have around generalization of group-alike methods in other PRs – instead of hardcoding output types we could try to elaborate on making them more generic.

[eugkhp]

@armanbilge yep! Should I close this one or we can merge it and I will work from there?

[eugkhp]

made another PR with proposed changes #4336