Open georgefst opened 2 years ago
Yes, good observation. Mon
is pretty ancient, dating back to 9b4450560018e6d2e876a6184737fa5a320998cd (before the Applicative
–Monad
proposal, IIRC). Now that Applicative
is a superclass of Monad
, there's not much reason to use Mon
over Ap
. I'd be happy to deprecate the former in favor of the latter.
Speaking of which, reducers
has its own cargo-culted version of Ap
here. Perhaps we should use the version from base
on sufficiently recent versions of GHC.
Isn't
Data.Semigroup.Monad.Mon
identical tobase
'sData.Monoid.Ap
?Either way, it would be worth a haddock comment clarifying the relationship, and why
Mon
exists (just for backwards compatiblity?).