X-Git-Url: http://lambda.jimpryor.net/git/gitweb.cgi?p=lambda.git;a=blobdiff_plain;f=topics%2F_week7_monads.mdwn;h=7d189e4dab6f44745624dfa6ad33228f1b9d43fe;hp=21ac2f59d098db062ed50f14ada225764fa211f5;hb=d198ef88b87d24fb20d2153a5c49be6d1285668e;hpb=5186270ce99d2ebe9e761defb1fbf63bd002d679
diff --git a/topics/_week7_monads.mdwn b/topics/_week7_monads.mdwn
index 21ac2f59..7d189e4d 100644
--- a/topics/_week7_monads.mdwn
+++ b/topics/_week7_monads.mdwn
@@ -1,3 +1,6 @@
+
+
+
Monads
======
@@ -5,13 +8,169 @@ The [[tradition in the functional programming
literature|https://wiki.haskell.org/Monad_tutorials_timeline]] is to
introduce monads using a metaphor: monads are spacesuits, monads are
monsters, monads are burritos. We're part of the backlash that
-prefers to say that monads are monads.
+prefers to say that monads are (Just) monads.
The closest we will come to metaphorical talk is to suggest that
-monadic types place objects inside of boxes, and that monads wrap and
-unwrap boxes to expose or enclose the objects inside of them. In any
-case, the emphasis will be on starting with the abstract structure of
-monads, followed by instances of monads from the philosophical and
+monadic types place objects inside of *boxes*, and that monads wrap
+and unwrap boxes to expose or enclose the objects inside of them. In
+any case, the emphasis will be on starting with the abstract structure
+of monads, followed by instances of monads from the philosophical and
linguistics literature.
-To
+## Box types: type expressions with one free type variable
+
+Recall that we've been using lower-case Greek letters
+α, β, γ, ...
as variables over types. We'll
+use `P`, `Q`, `R`, and `S` as metavariables over type schemas, where a
+type schema is a type expression that may or may not contain unbound
+type variables. For instance, we might have
+
+ P_1 â¡ Int
+ P_2 ⡠α -> α
+ P_3 â¡ âα. α -> α
+ P_4 â¡ âα. α -> β
+
+etc.
+
+A *box type* will be a type expression that contains exactly one free
+type variable. Some examples (using OCaml's type conventions):
+
+ α Maybe
+ α List
+ (α, P) Tree (assuming P contains no free type variables)
+ (α, α) Tree
+
+The idea is that whatever type the free type variable α might be,
+the boxed type will be a box that "contains" an object of type α.
+For instance, if `α List` is our box type, and α is the basic type
+Int, then in this context, `Int List` is the type of a boxed integer.
+
+We'll often write box types as a box containing the value of the free
+type variable. So if our box type is `α List`, and `α == Int`, we
+would write
+
+Int
+
+for the type of a boxed Int.
+
+## Kleisli arrows
+
+At the most general level, we'll talk about *Kleisli arrows*:
+
+P -> Q
+
+A Kleisli arrow is the type of a function from objects of type P to
+objects of type box Q, for some choice of type expressions P and Q.
+For instance, the following are arrows:
+
+Int -> Bool
+
+Int List -> Int List
+
+Note that the left-hand schema can itself be a boxed type. That is,
+if `α List` is our box type, we can write the second arrow as
+
+Int -> Int
+
+We'll need a number of classes of functions to help us maneuver in the
+presence of box types. We will want to define a different instance of
+each of these for whichever box type we're dealing with. (This will
+become clearly shortly.)
+
+mid (/εmaidεnt@tI/ aka unit, return, pure): P -> P
+
+map (/maep/): (P -> Q) -> P -> Q
+
+map2 (/m&ash;ptu/): (P -> Q -> R) -> P -> Q -> R
+
+mapply (/εm@plai/): P -> Q -> P -> Q
+
+mcompose (aka <=<): (Q -> R) -> (P -> Q) -> (P -> R)
+
+mbind (aka >>=): ( Q) -> (Q -> R) -> ( R)
+
+mflipcompose (aka >=>): (P -> Q) -> (Q -> R) -> (P -> R)
+
+mflipbind (aka =<<) ( Q) -> (Q -> R) -> ( R)
+
+mjoin: P -> P
+
+The managerie isn't quite as bewildering as you might suppose. For
+one thing, `mcompose` and `mbind` are interdefinable: u >=> k â¡
+\a. (ja >>= k)
.
+
+In most cases of interest, instances of these types will provide
+certain useful guarantees.
+
+* ***Mappable*** ("functors") At the most general level, box types are *Mappable*
+if there is a `map` function defined for that box type with the type given above.
+
+* ***MapNable*** ("applicatives") A Mappable box type is *MapNable*
+ if there are in addition `map2`, `mid`, and `mapply`. (With
+ `map2` in hand, `map3`, `map4`, ... `mapN` are easily definable.)
+
+* ***Monad*** ("composables") A MapNable box type is a *Monad* if there
+ is in addition an `mcompose` and a `join` such that `mid` is
+ a left and right identity for `mcompose`, and `mcompose` is
+ associative. That is, the following "laws" must hold:
+
+ mcompose mid k = k
+ mcompose k mid = k
+ mcompose (mcompose j k) l = mcompose j (mcompose k l)
+
+To take a trivial (but, as we will see, still useful) example,
+consider the identity box type Id: `α -> α`. So if α is type Bool,
+then a boxed α is ... a Bool. In terms of the box analogy, the
+Identity box type is a completly invisible box. With the following
+definitions
+
+ mid â¡ \p.p
+ mcompose â¡ \fgx.f(gx)
+
+Id is a monad. Here is a demonstration that the laws hold:
+
+ mcompose mid k == (\fgx.f(gx)) (\p.p) k
+ ~~> \x.(\p.p)(kx)
+ ~~> \x.kx
+ ~~> k
+ mcompose k mid == (\fgx.f(gx)) k (\p.p)
+ ~~> \x.k((\p.p)x)
+ ~~> \x.kx
+ ~~> k
+ mcompose (mcompose j k) l == mcompose ((\fgx.f(gx)) j k) l
+ ~~> mcompose (\x.j(kx)) l
+ == (\fgx.f(gx)) (\x.j(kx)) l
+ ~~> \x.(\x.j(kx))(lx)
+ ~~> \x.j(k(lx))
+ mcompose j (mcompose k l) == mcompose j ((\fgx.f(gx)) k l)
+ ~~> mcompose j (\x.k(lx))
+ == (\fgx.f(gx)) j (\x.k(lx))
+ ~~> \x.j((\x.k(lx)) x)
+ ~~> \x.j(k(lx))
+
+Id is the favorite monad of mimes everywhere.
+
+To take a slightly less trivial (and even more useful) example,
+consider the box type `α List`, with the following operations:
+
+ mid: α -> [α]
+ mid a = [a]
+
+ mcompose: (β -> [γ]) -> (α -> [β]) -> (α -> [γ])
+ mcompose f g a = concat (map f (g a))
+ = foldr (\b -> \gs -> (f b) ++ gs) [] (g a)
+ = [c | b <- g a, c <- f b]
+
+These three definitions are all equivalent. In words, `mcompose f g
+a` feeds the a (which has type α) to g, which returns a list of βs;
+each β in that list is fed to f, which returns a list of γs. The
+final result is the concatenation of those lists of γs.
+
+For example,
+
+ let f b = [b, b+1] in
+ let g a = [a*a, a+a] in
+ mcompose f g 7 = [49, 50, 14, 15]
+
+It is easy to see that these definitions obey the monad laws (see exercises).
+