At this point I think it is interesting to see what our cousins have done to address the problem of invisible partiality.

PureScript has the `Partial` typeclass¹. I suggest we ask their community how they feel about it, what has been their experience and if they could do it differently, what it would be.

---

¹ https://pursuit.purescript.org/packages/purescript-partial/3.0.0

I've been very much of two minds in this debate: On the one hand, having these constraints is very practically useful. On the other, what we're doing here is very un-Haskellish, in that we're letting operational concerns leak into a declarative property (a function's type). The reason we're doing this is another un-Haskellish thing -- partiality -- but that ship has sailed.

So, may I propose a slightly different way forward?

Instead of adding a HasCallStack constraint on these functions, add an IsPartial constraint. For example:

> head :: IsPartial => [a] -> a

This is slightly awkward, still, because IsPartial is a class-constraint-like-thing, but it has no parameter. But it has a few very nice properties:

* IsPartial is declarative: it describes a property of the function without worrying about its operation.

* If we think about the way constraints propagate, IsPartial has the right semantics: the caller of a partial function would itself become partial.

* We have some room in how we relate IsPartial to HasCallStack. We could say that IsPartial is just a synonym for HasCallStack (e.g. with type IsPartial = HasCallStack). But perhaps better would be to somehow give users control over whether they want the HasCallStack mechanism to be able to solve IsPartial constraints. Maybe some users would prefer not to be able to satisfy IsPartial constraints immediately, but instead to require an acknowledgement in their code that they're doing something partial. For example:

partialityIsOK :: String -> (IsPartial => r) -> r

elements xs = map (partialityIsOK "lists returned by `group` are always non-empty" head) (group xs)

The partialityIsOK function has a more involved type than I would like, but it's very usable in practice. Of course, such a thing only makes sense if IsPartial cannot automatically be satisfied. Getting this to work properly probably needs an extra language feature (maybe make IsPartial magically built-in?), but it might provide a declarative, yet operationally practical way forward here.

Richard

On Jun 6, 2021, at 12:49 PM, Dominic Steinitz <dominic@steinitz.org> wrote:

-1 for the reasons Henrik has listed

Dominic Steinitz
dominic@steinitz.org
http://idontgetoutmuch.org
Twitter: @idontgetoutmuch

On 5 Jun 2021, at 11:10, libraries-request@haskell.org wrote:

Re: RFC: Add HasCallStack constraint to partial Data.List
functions.

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

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