The documentation of Zip says an instance of Zip must satisfy the following Distributivity laws.
-- The distributivity laws of Zip f
align (zip xs ys) zs ≡ undistrThesePair <$> zip (align xs zs) (align ys zs)
distrPairThese <$> zip (align xs ys) zs ≡ align (zip xs zs) (zip ys zs)
zip (align xs ys) zs ≡ undistrPairThese <$> align (zip xs zs) (zip ys zs)
Each of the three equations must hold. Let me name each equation as Distr(1) - Distr(3).
-- Distr(1)
align (zip xs ys) zs ≡ undistrThesePair <$> zip (align xs zs) (align ys zs)
-- Distr(2)
distrPairThese <$> zip (align xs ys) zs ≡ align (zip xs zs) (zip ys zs)
-- Distr(3)
zip (align xs ys) zs ≡ undistrPairThese <$> align (zip xs zs) (zip ys zs)
But only Distr(2) is necessary among three, if I'm not mistaken. The rest follows from Distr(2).
Gist: https://gist.github.com/viercc/5b79efc5c3593a1b487a51f04c2bffca#file-distributive-simplify-md
I propose changes to the documentation stating this.
- Remove
Distr(1) from the law and add a note that it follows from Distr(2).
- Just remove
Distr(3). It is a direct consequence of undistrPairThese . distrPairThese = id.
- Probably there should be a comment on
undistrPairThese that it is a left inverse (but not the inverse) of distrPairThese.
The documentation of Zip says an instance of
Zipmust satisfy the following Distributivity laws.Each of the three equations must hold. Let me name each equation as
Distr(1)-Distr(3).But only
Distr(2)is necessary among three, if I'm not mistaken. The rest follows fromDistr(2).Gist: https://gist.github.com/viercc/5b79efc5c3593a1b487a51f04c2bffca#file-distributive-simplify-md
I propose changes to the documentation stating this.
Distr(1)from the law and add a note that it follows fromDistr(2).Distr(3). It is a direct consequence ofundistrPairThese . distrPairThese = id.undistrPairThesethat it is a left inverse (but not the inverse) ofdistrPairThese.