Jacques Carette
Jacques Carette
I had already created that tag (where you'd tagged me on the original version of this issue about discoverability).
I am definitely a big fan of offering both narrow and wide versions of APIs. Library developers are going to want to be as precise as possible, with as few...
I personally expect `Properties` to be 'wide'. I would prefer `IsX` to be fairly narrow. In both cases, if a structure X includes a particular property already, then any structure...
I'm personally on the side of preferring findability over minimalism, as long as it does not incur a (predictable) non-trivial maintenance burden. In more detail: the 'core' of the library...
> are you arguing that eg Data.Bool, as the 'user-facing' module [...] Indeed, I would accept @andreasabel 's lemma in `Data.Bool`. It could have a direct or indirect proof -...
One big benefit of inductive types is "no junk no confusion" to re-use Goguen's nice phrasing. So if "no confusion" is not cheap, there is definitely a problem to be...
Note that I'm away at a conference right now, so won't be able to review in detail for another few days. Because PRED is essentially isomorphic to 'containers' which is...
Thanks for catching that I missed the orientation reversal. And yes, it probably is worth doing `Fam` instead. Do you know what that means @AJChapman ?
Maybe we'll end up with a v2.4 after all? Might be a small bump though.
As long as worthwhile, not-really-breaking improvements keep coming in, we should release them.