I appreciate this article. It's clear, easy to understand, and uses theory to propose practical improvements to FP.
Some minor nitpicks:
> In total functional programming ⊥ does not exist.
That doesn't seem right. In total FP, we can still define a ⊥ type that has no values. I think the problem here is that the author uses ⊥ to indicate both a type and a value of that type (e.g. f ⊥ = ⊥), which is confusing.
> 0 / 0 = 0
Surely it would be better to use a Maybe type for this instead?
0 / 1 = Just 0
0 / 0 = Nothing
The same approach also addresses the difficulty of hd:
hd :: List a -> Maybe a
hd (Cons a x) = Just a
hd Nil = Nothing
F# has this function, and calls it "tryHead", rather than just "head".
Some minor nitpicks:
> In total functional programming ⊥ does not exist.
That doesn't seem right. In total FP, we can still define a ⊥ type that has no values. I think the problem here is that the author uses ⊥ to indicate both a type and a value of that type (e.g. f ⊥ = ⊥), which is confusing.
> 0 / 0 = 0
Surely it would be better to use a Maybe type for this instead?
The same approach also addresses the difficulty of hd: F# has this function, and calls it "tryHead", rather than just "head".