Closed nfrisby closed 1 week ago
For the record, I am in favour of (<>) = (.&&.)
as the least surprising choice. Another option is having newtypes like All a
and Any a
that are Testable
if a
is Testable
. That way nobody would be upset by the choice and the choice of instance would be crystal clear.
This Issue is complete once the library includes these instances.
I'm presuming that conjunction is the most useful and unsurprising monoidal semantics.
I think even just closing this Issue with an explanation for why we don't want this instance would also be quite useful as documentation. Thanks.
(I was surprised that I could not find a pre-existing Issue for this -- I apologize if I just missed it. I was also surprised to not find discussions of this elsewhere; it seemed like "
QuickCheck
properties of monoids" was masking all my searches.)