Let's just pause and consider what is already available on hackage today for these situations:

In my constraints package I have a class named `Lifting`, which provides.

class Lifting p f where
    lifting :: p a :- p (f a)

Lifting Eq, Lifting Monad, Lifting Semigroup, Lifting (MonadReader e), etc. are then able to be handled all uniformly.

It is, alas, somewhat annoying to use, as you need to use `\\ lifting` with a scoped type variable signature to get the instance in scope

The currrent Eq1 is a somewhat more powerful claim though, since you can supply the equality for its argument without needing functoriality in f. This is both good and bad. It means you can't just write `instance Eq1 f` and let default methods take over, but it does mean Eq1 f works in more situations if you put in the work or use generics to generate it automatically.

http://hackage.haskell.org/package/constraints-0.8/docs/Data-Constraint-Lifting.html

For the rank-2 situation, I also have `Forall` and `ForallF` which provides the ability to talk about the quantified form.

ForallF Eq f is defined by a fancy skolem type family trick and comes with

instF :: forall p f a. ForallF p f :- p (f a)

This covers the rank-2 situation today pretty well, even if you have to use `\\ instF` or what have you to get the instance in scope.

I don't however, have something in a "mainstream" package for that third form mentioned above, the 'Functor'-like form, but I do have classes in semgroupoids for Alt, Plus, etc. covering the particular semigroup/monoid-like cases. 

Finally, going very far off the beaten and well-supported path, in `hask`, I have code for talking about entailment in the category of constraints, but like the above two tricks, it requires the user to explicitly bring the instance into scope from an `Eq a |- Eq (f a)` constraint or the like, and the more general form of `|-` lifts into not just Constraint, but k -> Constraint, and combines with Lim functor to provide quantified entailment. This doesn't compromise the thinness of the category of constraints. I'd love to see compiler support for this, eliminating the need for the \\ nonsense above, but it'd be a fair bit of work!

-Edward

On Sat, Oct 1, 2016 at 2:10 PM, Mario Blažević <blamario@ciktel.net> wrote:

CC-ing the Café on class naming...

On 2016-10-01 04:07 AM, Edward Kmett wrote:

I'm somewhat weakly against these, simply because they haven't seen
broad adoption in the wild in any of the attempts to introduce them
elsewhere, and they don't quite fit the naming convention of the other
Foo1 classes in Data.Functor.Classes

Eq1 f says more or less that Eq a => Eq (f a).

Semigroup1 in your proposal makes a stronger claim. Semgiroup1 f  is
saying forall a. (f a) is a semigroup parametrically. Both of these
constructions could be useful, but they ARE different constructions.

     The standard fully parametric classes like Functor and Monad have no suffix at all. It makes sense to reserve the suffix "1" for non-parametric lifting classes. Can you suggest a different naming scheme for parametric classes of a higher order?

     I'm also guilty of abusing the suffix "1", at least provisionally, but these are different beasts yet again:

   -- | Equivalent of 'Functor' for rank 2 data types
   class Functor1 g where
   fmap1 :: (forall a. p a -> q a) -> g p -> g q

https://github.com/blamario/grampa/blob/master/Text/Grampa/Classes.hs

     What would be a proper suffix here? I guess Functor2 would make sense, for a rank-2 type?

If folks had actually been using, say, the Plus and Alt classes from
semigroupoids or the like more or less at all pretty much anywhere, I
could maybe argue towards bringing them up towards base, but I've seen
almost zero adoption of the ideas over multiple years -- and these
represent yet _another_ point in the design space where we talk about
semigroupal and monoidal structures where f is a Functor instead. =/

Many points in the design space, and little demonstrated will for
adoption seems to steers me to think that the community isn't ready to
pick one and enshrine it some place central yet.

Overall, -1.

-Edward

On Fri, Sep 30, 2016 at 7:25 PM, David Feuer <david.feuer@gmail.com
<mailto:david.feuer@gmail.com>> wrote:

    I've been playing around with the idea of writing Haskell 2010
    type classes for finite sequences and non-empty sequences,
    somewhat similar to Michael Snoyman's Sequence class in
    mono-traversable. These are naturally based on Monoid1 and
    Semigroup1, which I think belong in base.

    class Semigroup1 f where
      (<<>>) :: f a -> f a -> f a
    class Semigroup1 f => Monoid1 f where
      mempty1 :: f a

    Then I can write

    class (Monoid1 t, Traversable t) => Sequence t where
      singleton :: a -> t a
      -- and other less-critical methods

    class (Semigroup1 t, Traversable1 t) => NESequence where
      singleton1 :: a -> t a
      -- etc.

    I can, of course, just write my own, but I don't think I'm the
    only one using such.


    _______________________________________________
    Libraries mailing list
    Libraries@haskell.org <mailto:Libraries@haskell.org>
    http://mail.haskell.org/cgi-bin/mailman/listinfo/libraries
    <http://mail.haskell.org/cgi-bin/mailman/listinfo/libraries>




_______________________________________________
Libraries mailing list
Libraries@haskell.org
http://mail.haskell.org/cgi-bin/mailman/listinfo/libraries

_______________________________________________
Haskell-Cafe mailing list
To (un)subscribe, modify options or view archives go to:
http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
Only members subscribed via the mailman list are allowed to post.

_______________________________________________
Haskell-Cafe mailing list
To (un)subscribe, modify options or view archives go to:
http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
Only members subscribed via the mailman list are allowed to post.