To explain these concepts in more detail, you want a Sigma type (n :: Nat, Vect (n :: Nat) a) But we don’t have those in Haskell. Instead you can have a second class existential type by writing it as a GADT: data SomeVect a where SomeVect :: (n :: Nat) -> Vect (n :: Nat) a -> SomeNat a -- ^ This is not yet valid Haskell! But we don’t have dependent types either! Instead there is a singleton type SNat :: Nat -> Type that does a similar job. It’s called a “singleton” because there is ...
