On 15 Feb 2012, at 01:03, Jeff Shaw wrote:
If we want to argue about maybeTo and the Fairbairn threshold, probably we should argue about whether to include maybeToList, which is less general than maybeTo.
and only a fraction shorter than foldMap pure or, as it is in the Epigram codebase trail :: (Applicative f, Foldable t, Monoid (f a)) => t a -> f a trail = foldMap pure "a trail told by an idiom..." being the joke at the time. It would do listToMaybe as well, if the Monoid instance for Maybe captured the prioritized choice semantics. It would be interesting to apply the Fairbairn threshold retrospectively and see what could be cheaply replaced by suitable deployment of Applicative/Traversable/Foldable/Monoid and (crucially) Control.Newtype. Instead of remembering lots of names of functions, get types to remember things for you! Or something. Gotta run Conor