Proposal: Num instance for (a -> b)

[Jon Fairbairn]

Daniel Cartwright <chessai1996 at gmail.com> writes:

> ANum seems to be just Data.Monoid.Ap.
> Also, I can see not wanting to worsen the error messages, though it is
> worth pointing out that we already have a Monoid instance with the same
> semantics, and a similar potential for confusing error messages.

The existence of a previous bad decision doesn’t seem to me to
be a good reason to make another one.

-- 
Jón Fairbairn                                 Jon.Fairbairn at cl.cam.ac.uk
http://www.chaos.org.uk/~jf/Stuff-I-dont-want.html  (updated 2014-04-05)