Wrap output of splitIntoCoprimes into newtype

[Bodigrim]

Currently splitIntoCoprimes guarantees certain properties of its output, but still returns a plain [(a, b)]. Such return type is easily corruptible by other functions. It would be much better to return something like

newtype Coprimes a b = Coprimes (Map a b)

The interface should not expose Coprimes constructor outside. Here is a quick draft of API:

instance Semigroup (Coprimes a b) where ...

instance Monoid (Coprimes a b) where ...

insert :: a -> b -> Coprimes a b -> Coprimes a b 

toList :: Coprimes a b -> [(a, b)]

---

Another vaguely related idea is to keep coprimes, which are known to be prime, separately. This can potentially speed up certain computations, especially with Prefactored. Like

data Coprimes a b = Coprimes (Map (Prime a) b) (Map a b)

[cartazio]

Style wise I like to use record fields to document what different positions in a tuple mean. (One of my fave party tricks Is showing that you can mix record and fast syntax. Though that’s not relevant here ;) )

…

On Tue, Jan 2, 2018 at 5:16 PM Bodigrim ***@***.***> wrote: Currently splitIntoCoprimes <https://github.com/cartazio/arithmoi/blob/master/Math/NumberTheory/GCD.hs#L267> guarantees certain properties of its output, but still returns a plain [(a, b)]. Such return type is easily corruptible by other functions. It would be much better to return something like newtype Coprimes a b = Coprimes (Map a b) The interface should not expose Coprimes constructor outside. Here is a quick draft of API: instance Semigroup (Coprimes a b) where ... instance Monoid (Coprimes a b) where ... insert :: a -> b -> Coprimes a b -> Coprimes a b toList :: Coprimes a b -> [(a, b)] ------------------------------ Another vaguely related idea is to keep coprimes, which are known to be prime, separately. This can potentially speed up certain computations, especially with Prefactored. Like data Coprimes a b = Coprimes (Map (Prime a) b) (Map a b) — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#89>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAAQwpoXD9xxn5LBbmRijtJQ7wlrCC1Gks5tGqrPgaJpZM4RRJFB> .