## Thursday, 24 December 2009

## Sunday, 29 November 2009

## Tuesday, 27 October 2009

### Cantors Diagonalization Proof in Coq

Axiom b : (nat -> nat) -> nat.

Axiom b' : nat -> (nat -> nat).

Axiom bijection : forall f, b' (b f) = f.

Axiom bijection' : forall x, b (b' x) = x.

Definition Stream A := nat -> A.

Definition Map {A B} (f : A -> B) : Stream A -> Stream B :=

fun sa =>

fun i =>

f (sa i).

Definition l : Stream (nat -> nat) -> Stream nat :=

Map b.

Definition l' : Stream nat -> Stream (nat -> nat) :=

Map b'.

Definition isnt x :=

match x with

| O => S O

| S _ => O

end.

Theorem isn't_isnt {x} : isnt x <> x.

destruct x; discriminate.

Qed.

Definition Diagonal {A} (s : Stream (Stream A)) : Stream A:=

fun i => s i i.

Definition Diagonalization (sequences : Stream (nat -> nat)) : Stream nat :=

Map isnt (Diagonal sequences).

Theorem isn't_in {sequences} : forall i, Diagonalization sequences <> sequences i.

intros; intro.

pose (f_equal (fun f => f i) H) as H'.

unfold Diagonalization in H'.

unfold Diagonal in H'.

unfold Map in H'.

revert H'.

apply isn't_isnt.

Qed.

Definition pink_elephant : Stream nat := fun i => i.

Definition sequence_enumeration : Stream (nat -> nat) := l' pink_elephant.

Theorem every_sequence : forall s, exists i, s = sequence_enumeration i.

intro s; exists (b s).

unfold sequence_enumeration.

unfold pink_elephant.

unfold l'.

unfold Map.

rewrite bijection.

reflexivity.

Qed.

Theorem Cantor : False.

destruct (every_sequence (Diagonalization sequence_enumeration)) as [i iPrf].

pose (@isn't_in sequence_enumeration i).

contradiction.

Qed.

## Saturday, 24 October 2009

### Short Note on Semantics

With all curious derivations that come from the forests of To Mock a Mockingbird, and in the intrests of deepening understanding of type theory (in particular CIC) we considered axiomatizing the SK calculus in Coq.

Module Type SK.

(* axiomatization of the theory *)

Parameter SK : Set.

Parameter app : SK -> SK -> SK.

Infix "*" := app.

Parameter S : SK.

Parameter K : SK.

Axiom S_prop : forall a b c, S * a * b * c = (a * c) * (b * c).

Axiom K_prop : forall a b, K * a * b = a.

End SK.

It is not defined as an inductive type (as you might usually do) because there is no normal form for SK that we can compute terms into inside Coq (we would have to equate diverging terms with an "omega" symbol for a start, which is well known to be a partial operation). It would be possible to use a non-normal forma and quotient it out with a special (undecidable) equivalence relation, but we do not this.. Instead the axioms of the theory are given, One immediately notices that there is a trivial model of these axioms:

Module SK_trivial_model : SK.

(* one model (giving non-contradiction of the theory) *)

Definition SK := unit.

Definition app := fun (_ _ : unit) => tt.

Infix "*" := app.

Definition S := tt.

Definition K := tt.

Theorem S_prop : forall a b c, S * a * b * c = (a * c) * (b * c).

reflexivity.

Qed.

Theorem K_prop : forall a b, K * a * b = a.

destruct a; reflexivity.

Qed.

End SK_trivial_model.

(This implementation mostly serves to show the syntax and so on of how to construct a model of a theory inside Coq).

Anyway, we eliminate trivial models by saying that S <> K.

Axiom SK_prop : S <> K.

We could now develop the theory of combinator calculus, prove things like S <> I and so forth.

Module SK_theory <: SK.

Include Type SK.

(* Development of the theory of SK calculus, valid for all models *)

Ltac sk := repeat (rewrite S_prop || rewrite K_prop).

Definition I := S * K * S.

Theorem I_prop : forall x, I * x = x.

intros; unfold I; sk; reflexivity.

Qed.

Ltac ski := repeat (sk || rewrite I_prop).

Theorem IK : I <> K.

Definition F x y b := b * K * I * x * y.

intro Eq; generalize (f_equal (F K S) Eq); unfold F; ski.

apply SK_prop.

Qed.

Definition U := S * I * I.

Theorem U_prop : forall x, U * x = x * x.

intros; unfold U; ski; reflexivity.

Qed.

Ltac skiu := repeat (ski; rewrite U_prop).

End SK_theory.

As fun as that would be, I don't believe you could ever reach a contradiction by assuming the SK theory. Yet no model exists inside Coq!

This is a rather confusing and troubling state, but Russell O'Connor mentioned to me the possibility of giving an interpretation of CIC into a theory that has a Halting-Oracle and that solves it! This is surely a valid interpretation of CIC, it's just got this SK theory added in with the usual stuff - the halting oracle would be used to normalize SK terms and prove satisfy our theory. So despite the non-existence of a model of SK inside Coq: It appears we can reason about **the** SK calculus directly. (Rather than a halting oracle you might have considered a special equivalence relation being interpreted for the congruence of SK objects but lets pretend Coq has Axiom K so that need not be considered).

This is all very interesting to me because it seems to suggest mathematics existing outside of the formal theory we work in or study. Something I never believed in - It's not completely clear if this is profound or if I am just going soft.

## Tuesday, 25 August 2009

## Sunday, 26 April 2009

### Strongly Specified Parser Combinators

The computational aspect of it is the same old parser monad, type Parser s a = [s] -> [(s, [a])], which takes a list of tokens [s] to a (possibly empty) list (here is where backtracking/nondeterminism comes into it) of parses paired with the rest of the text. It's a monad and also a monad plus.

On the specification end, we can put a precondition on the input string (for example, you might say the input length is <= some value, for ensuring termination of a recursive parser) and also a post condition comparing the input with the parsed value and the remaining output.

Definition Pre := list s -> Prop.

Definition Post (t : Set) := list s -> t -> list s -> Prop.

Program Definition Parser (pre : Pre) (t : Set) (post : Post t) : Set :=

forall i : { t : list s | pre t }, list ({ (x, r) : t * list s | post i x r }).

in Haskell we might define:

m >>= f = \i -> concatMap (uncurry f) (m i)

And in Coq, using Program we have roughly the same thing. Except that one has to apply a 'noncomputational_map' to fudge the proofs paired up with the list elements.

Program Definition Bind (a b : Set) P1 P2 Q1 Q2

(m : Parser P1 a Q1)

(f : (forall x : a, Parser (P2 x) b (Q2 x))) :

Parser (fun i => P1 i /\ forall x o, Q1 i x o -> P2 x o)

b

(fun i x' o' => exists x o, Q1 i x o /\ Q2 x o x' o') :=

fun i => @flat_map ({ (x, o) : a * list s | Q1 i x o }) _

(fun xo => match xo with (x,o) => noncomputational_map _ _ _ _ (f x o) end)

(m i).

Seeing as noncomputational map doesn't do anything, we prove a theorem expressing that as justification for extracting it out as the identity function (rather it being equivalent to map id traversing the whole list).

Theorem noncomputational_map_identity :

forall l,

map (@proj1_sig _ _) l = map (@proj1_sig _ _) (noncomputational_map l).

Extract Inlined Constant noncomputational_map => "id".

Another nice parser combinator is the fixed point of a parser:

Program Definition Fix t {P Q} :

(forall i : { t : list s | P t },

Parser (fun i' => length i' < length i /\ P i') t Q ->

list ({ (x, o) : t * list s | Q i x o })) ->

Parser P t Q :=

fun Rec =>

well_founded_induction

(well_founded_ltof ({ i : list s | P i }) (fun i => length i))

(fun i => list ({ (x, o) : t * list s | Q i x o }))

(fun x Rec' => Rec x (fun i => Rec' i _)).

It packages up well founded recursion on the size of the input string, so that any non-left recursive parsers should be easily defined.

Enough of the heavy machinary! A simple example of putting thing into work now:

<par> ::= <epsilon> | '(' <par> ')' <par>

To parse this we define a type of tokens and abstract syntax of parsing derivations - then relate them with a function:

Inductive token := open | close.

Inductive par := epsilon : par | wrappend : par -> par -> par.

Fixpoint print (p : par) : list token :=

match p with

| epsilon => nil

| wrappend m n => (open::nil) ++ print m ++ (close::nil) ++ print n

end.

Now Program lets us define the par parser as the fixed point of the sum of epsilon and wrapped recursions:

Program Definition par_parser : Parser token (fun _ => True) par (fun x p y => x = print p ++ y /\ length y <= length x) :=

Fix token par (fun i parRec =>

Plus _ _ (fun i' => i = i') _

(fudge_pre_and_post_conditions _ _ _

(Return epsilon))

(fudge_pre_and_post_conditions _ _ _

(Symbol token eq_token_dec open >>= fun _ =>

parRec >>= fun m =>

Symbol token eq_token_dec close >>= fun _ =>

parRec >>= fun n =>

Return (wrappend m n)))

i).

Here are the actual scripts http://github.com/odge/parseq/tree/master

## Friday, 27 March 2009

### Type Checker

import Control.Monad

import Control.Monad.Error

import Control.Monad.Identity

import Data.Maybe

import Data.Char

import Text.ParserCombinators.ReadP

-- Proof irrelevance is implemented by:

--

-- erase(Gamma) |- phi : P

-- ----------------------- Prf

-- Gamma |- []phi : Prf(P)

--

-- []phi =def []psi : Prf(_)

--

-- Allow eliminations like P(x) -> Prf(x = y) -> P(y)

-- along the same schema as Coqs Prop sort, this gives us Axiom K.

-- * I am not a number: I am a free variable

-- * Verifying a Semantic beta-eta-Conversion Test for Martin-Löf Type Theory

-- * Intensionality, extensionality, and proof irrelevance in modal type theory

-- Syntax

type Ref = Int

newtype Scope a = Scope a

deriving Show

data Name

= Local Ref

| Global String

deriving (Eq, Show)

data T

= Type

| Pi T (Scope T)

| Prf T

| Lam T (Scope T)

| Square T

| App T T

| Def String

| Bound Ref

| Free Name

deriving Show

abstract x expr = Scope (rip 0 expr) where

rip i Type = Type

rip i (Pi tau (Scope sigma)) = Pi (rip i tau) (Scope (rip (i+1) sigma))

rip i (Prf p) = Prf (rip i p)

rip i (Lam tau (Scope sigma)) = Lam (rip i tau) (Scope (rip (i+1) sigma))

rip i (Square s) = Square (rip i s)

rip i (App m n) = App (rip i m) (rip i n)

rip i (Def d) = Def d

rip i (Bound j) = Bound j

rip i (Free j)

| x == j = Bound i

| otherwise = Free j

instantiate y (Scope expr) = shove 0 expr where

shove i Type = Type

shove i (Pi tau (Scope sigma)) = Pi (shove i tau) (Scope (shove (i+1) sigma))

shove i (Prf p) = Prf (shove i p)

shove i (Lam tau (Scope sigma)) = Lam (shove i tau) (Scope (shove (i+1) sigma))

shove i (Square s) = Square (shove i s)

shove i (App m n) = App (shove i m) (shove i n)

shove i (Def d) = Def d

shove i (Bound j)

| i == j = y

| otherwise = Bound j

shove i (Free j) = Free j

-- Parsing

-- *

-- /\(x:T),x

-- Prf(P)

-- \(x:T),x

-- []P

-- T T

-- x

parseTerm = readP_to_S pTerm

pTerm = pPi +++ pLam +++

do a <- skipSpaces >> pSimpleUnit ; as <- many (skipSpaces >> pUnit) ; skipSpaces ; return (foldl1 App (a:as))

pUnit = pType +++ pPi +++ pPrf +++ pLam +++ pSquare +++ pFree +++ parens pTerm

pSimpleUnit = pType +++ pPrf +++ pSquare +++ pFree +++ parens pTerm

pType = do string "*" ; return Type

pPi = do string "/\\" ; (x, t) <- binder ; string "," ; b <- pTerm ; return (Pi t (abstract x b))

pPrf = do string "Prf" ; m <- parens pTerm ; return (Prf m)

pLam = do string "\\" ; (x, t) <- binder ; string "," ; b <- pTerm ; return (Lam t (abstract x b))

pSquare = do string "[]" ; m <- pTerm ; return (Square m)

pFree = do a <- name ; guard (a /= "Prf") ; return (Free (Global a))

name = munch1 isAlpha

binder = b +++ parens b where

b = do a <- name ; string ":" ; x <- pTerm ; return (Global a,x)

parens p = do string "(" ; r <- p ; string ")" ; return r

-- bug 1: (\x:*,\y:*,x y)

-- parses as: (\x:*,(\y:*,(x y)))

-- or: (\x:*,(\y:*,x) y)

-- or: ((\x:*,\y:*,x) y)

-- fix: introduced 'simple' units and changed pTerm

-- from being applications of [unit unit ...] to [simpleUnit unit ...]

-- Typing

data V

= VType

| VPi V (V -> V)

| VPrf V

| VLam (V -> V)

| VSquare V

| VNeutral N

data N

= NApp N V

| NFree Name

VLam f % x = f x

VNeutral n % x = VNeutral (NApp n x)

_ % _ = error "Malformed application"

eval delta xi Type = VType

eval delta xi (Pi tau (Scope sigma)) = VPi (eval delta xi tau) (\x -> eval (x:delta) xi sigma)

eval delta xi (Prf p) = VPrf (eval delta xi p)

eval delta xi (Lam _ (Scope f)) = VLam (\x -> eval (x:delta) xi f)

eval delta xi (Square s) = VSquare (eval delta xi s)

eval delta xi (App m n) = eval delta xi m % eval delta xi n

eval delta xi (Def d) = fromJust (lookup d xi)

eval delta xi (Bound i) = delta !! i

eval delta xi (Free j) = VNeutral (NFree j)

quote q VType = Type

quote q (VPi tau sigmaF) = Pi (quote q tau) (abstract (Local q) (quote (q+1) (sigmaF (VNeutral (NFree (Local q))))))

quote q (VPrf p) = Prf (quote q p)

quote q (VLam f) = Lam Type (abstract (Local q) (quote (q+1) (f (VNeutral (NFree (Local q))))))

quote q (VSquare s) = Square (quote q s)

quote q (VNeutral n) = nquote q n

nquote q (NApp n v) = App (nquote q n) (quote q v)

nquote q (NFree j) = Free j

abstractV i VType _ = VType

abstractV i (VPi tau sigmaF) o = VPi (abstractV i tau o) (\e -> abstractV i (sigmaF e) o)

abstractV i (VPrf p) x = VPrf (abstractV i p x)

abstractV i (VLam sigmaF) o = VLam (\e -> abstractV i (sigmaF e) o)

abstractV i (VSquare s) x = VSquare (abstractV i s x)

abstractV i (VNeutral n) x = abstractN i n x

abstractN i (NApp n m) x = abstractN i n x % abstractV i m x

abstractN i (NFree j) x

| i == j = x

| otherwise = VNeutral (NFree j)

type Infer = ErrorT String Identity

runInfer :: Infer o -> Either String o

runInfer m = runIdentity (runErrorT m)

conv q gamma VType VType = True

conv q gamma (VPi tau sigmaF) (VPi tau' sigmaF') =

conv q gamma tau tau' && conv (q+1) ((Local q, tau):gamma) (sigmaF x) (sigmaF' x) where

x = VNeutral (NFree (Local q))

conv q gamma (VPrf p) (VPrf p') = conv q gamma p p'

conv q gamma (VNeutral n) (VNeutral m) = maybe False (const True) (convN q gamma n m)

conv q gamma _ _ = False

convV q gamma f g (VPi tau sigmaF) = convV (q+1) ((Local q, tau):gamma) (f % x) (g % x) (sigmaF x)

where x = VNeutral (NFree (Local q))

convV q gamma _ _ (VPrf _) = True

convV q gamma (VNeutral n) (VNeutral m) _ = maybe False (const True) (convN q gamma n m)

convV q gamma t s VType = conv q gamma t s

convN q gamma (NFree i) (NFree j)

| i == j = lookup i gamma

convN q gamma (NApp f x) (NApp g y) = do

VPi tau sigmaF <- convN q gamma f g

case convV q gamma x y tau of

True -> Just (sigmaF x)

False -> Nothing

convN q gamma n m = do

case (convN q gamma n n, convN q gamma m m) of

(Just (VPrf p), Just (VPrf p')) | conv q gamma (VPrf p) (VPrf p') -> Just (VPrf p)

_ -> Nothing

erase = map (\(q,tau) -> (q,eraseV tau))

eraseV (VPrf p) = eraseV p

eraseV sigma = sigma

infer :: Ref -> [(Name, V)] -> [(String, V)] -> T -> Infer V

infer q gamma xi Type = return VType

infer q gamma xi (Pi tau scopeSigma) = do

VType <- infer q gamma xi tau

VType <- infer (q+1) ((Local q, VType):gamma) xi (instantiate (Free (Local q)) scopeSigma)

return VType

infer q gamma xi (Prf p) = do

VType <- infer q gamma xi p

return VType

infer q gamma xi (Lam tau scopeM) = do

VType <- infer q gamma xi tau

tau' <- return $ eval [] xi tau

sigmaF <- infer (q+1) ((Local q, tau'):gamma) xi (instantiate (Free (Local q)) scopeM)

pi <- return $ VPi tau' (abstractV (Local q) sigmaF)

return pi

infer q gamma xi (Square s) = do

p <- infer q (erase gamma) xi s -- How do local definitions relate to erasure?

return (VPrf p)

infer q gamma xi (App m n) = do

pi <- infer q gamma xi m

case pi of

VPi sigma tauF -> do

sigma' <- infer q gamma xi n

case conv q gamma sigma sigma' of

True -> return (tauF (eval [] xi n))

False -> throwError (unlines [ "Type error in application: "

, " expected: " ++ show (quote q sigma)

, " received: " ++ show (quote q sigma')

])

_ -> throwError "Nonfunctional application"

infer q gamma xi (Def d) = do

infer q gamma xi (Free (Global d))

infer q gamma xi (Bound _) = do

throwError "Encountered a bound variable"

infer q gamma xi (Free j) = do

case lookup j gamma of

Just tau -> return tau

Nothing -> case j of

Local _ -> throwError "Free local variable"

Global j -> throwError ("Free global variable: " ++ j)

inferType gamma xi t = case runInfer (infer q gamma xi t) of

Left err -> Left err

Right tau -> Right (quote q tau)

where q = length gamma + 1

checkType gamma xi t tau = case runInfer (infer q gamma xi t) of

Left err -> Left err

Right tau' -> Right (conv q gamma tau tau')

where q = length gamma + 1

unstickDefinitions xi Type = Type

unstickDefinitions xi (Pi t (Scope b)) = Pi (unstickDefinitions xi t) (Scope (unstickDefinitions xi b))

unstickDefinitions xi (Prf p) = Prf (unstickDefinitions xi p)

unstickDefinitions xi (Lam t (Scope b)) = Lam (unstickDefinitions xi t) (Scope (unstickDefinitions xi b))

unstickDefinitions xi (Square p) = Square (unstickDefinitions xi p)

unstickDefinitions xi (App m n) = App (unstickDefinitions xi m) (unstickDefinitions xi n)

unstickDefinitions xi (Def d) = Def d

unstickDefinitions xi (Bound j) = Bound j

unstickDefinitions xi (Free (Local j)) = Free (Local j)

unstickDefinitions xi (Free (Global g)) =

case lookup g xi of

Nothing -> Free (Global g)

Just _ -> Def g

-- TODO: that was silly, find a sensible fold for T

-- User interfacing

repl gamma xi = do

putStr "> "

line <- getLine

case parseTerm line of

[] -> putStrLn "No parse"

[(t,"")] -> case inferType gamma xi (unstickDefinitions xi t) of

Left err -> putStrLn err

Right ty -> print ty

[(_,stuff)] -> putStrLn ("extra stuff at end of line: " ++ stuff)

ambiguous -> putStrLn ("ambiguous parse: " ++ show (length ambiguous) ++ " alternatives")

repl gamma xi

-- Examples/Toys

-- nat : *; Z : nat; S : nat -> nat

nat =

[ (Global "nat", VType)

, (Global "Z", zType)

, (Global "S", sType)

, (Global "natElim", natElimType)

]

natDefs =

[ ("natElim", natElimV)

]

zType = VNeutral (NFree (Global "nat"))

sType = VPi (VNeutral (NFree (Global "nat"))) (\_ -> VNeutral (NFree (Global "nat")))

natElimType = VPi (VPi (VNeutral (nfree "nat")) (\_ -> VType)) (\p -> -- forall P : nat -> Type,

VPi (p % (VNeutral (nfree "Z"))) (\_ -> -- P 0 ->

VPi (VPi (VNeutral (nfree "nat")) (\n -> -- (forall n : nat,

VPi (p % n) (\_ -> (p % (VNeutral (nfree "S") % n))))) (\_ -> -- P n -> P (S n)) ->

VPi (VNeutral (nfree "nat")) (\n -> p % n)))) -- forall n : nat, P n

where nfree = NFree . Global

natElimV = VLam (\motive -> VLam (\zero -> VLam (\succ -> VLam (\n -> natElim motive zero succ n))))

natElim motive zero succ (VNeutral (NFree (Global "Z"))) = zero

natElim motive zero succ (VNeutral (NApp (NFree (Global "S")) m)) = succ % m % natElim motive zero succ m

natElim motive zero succ stuck = VNeutral (NFree (Global "natElim")) % motive % zero % succ % stuck

-- add : Pi (Free (Global "nat")) (Scope (Pi (Free (Global "nat")) (Scope (Free (Global "nat")))))

-- := (\(x:nat),\(y:nat),natElim (\(n:nat),nat) x (\(T:nat),S) y)

add =

[ (Global "add", addType)

]

addDefs =

[ ("add", addValue)

]

addType = VPi (VNeutral (NFree (Global "nat"))) (\_ ->

VPi (VNeutral (NFree (Global "nat"))) (\_ ->

VNeutral (NFree (Global "nat"))))

addValue = VLam (\x -> VLam (\y ->

natElimV % (VLam (\_ -> VNeutral (NFree (Global "nat")))) % y % (VLam (\_ -> VNeutral (NFree (Global "S")))) % x))

-- eq : /\T:*, T -> T -> *; refl : /\T:*,/\t:T, eq T t t

eq =

[ (Global "eq", VPi VType (\t -> VPi t (\_ -> VPi t (\_ -> VType))))

, (Global "refl", reflType)

, (Global "eqElim", eqElimType)

, (Global "eqPrfElim", eqPrfElimType) -- eq has only one constructor so allow Proof eliminations

]

eqDefs =

[ ("eqElim", eqElimV)

, ("eqPrfElim", eqElimV)

]

reflType = VPi VType (\t -> VPi t (\t' -> VNeutral (NFree (Global "eq")) % t % t' % t'))

eqElimType = VPi VType (\a -> VPi a (\x -> VPi (VPi a (\_ -> VType)) (\p -> -- forall (A : Type) (x : A) (P : A -> Type),

VPi (p % x) (\_ -> VPi a (\y -> -- P x -> forall y : A,

VPi (VNeutral (NFree (Global "eq")) % a % x % y) (\_ -> p % y)))))) -- x = y -> P y

eqPrfElimType = VPi VType (\a -> VPi a (\x -> VPi (VPi a (\_ -> VType)) (\p -> -- forall (A : Type) (x : A) (P : A -> Type),

VPi (p % x) (\_ -> VPi a (\y -> -- P x -> forall y : A,

VPi (VPrf (VNeutral (NFree (Global "eq")) % a % x % y)) (\_ -> p % y)))))) -- Prf(x = y) -> P y

eqElimV = VLam (\a -> VLam (\x -> VLam (\motive -> VLam (\pX -> VLam (\y -> VLam (\eqPrf -> eqElim pX eqPrf))))))

eqElim pX (VNeutral (NApp (NFree (Global "refl")) _)) = pX

eqElim _ _ = error "not implemented"

-- Example session:

--

-- Main> repl (nat ++ add ++ eq) (natDefs ++ addDefs ++ eqDefs)

--

-- -- join for the Prf monad:

--

-- > \P:*,\phi:Prf(Prf(P)),[]phi

-- Pi Type (Scope (Pi (Prf (Prf (Bound 0))) (Scope (Prf (Bound 1)))))

--

-- -- proof of eq symmetry:

--

-- > (\T:*,\a:T,\b:T,\e:eq T a b,eqElim T a (\x:T,eq T x a) (refl T a) b e)

-- Pi Type (Scope (Pi (Bound 0) (Scope (Pi (Bound 1) (Scope (Pi (App (App (App (Free (Global "eq")) (Bound 2)) (Bound 1)) (Bound 0)) (Scope (App (App (App (Free (Global "eq")) (Bound 3)) (Bound 1)) (Bound 2)))))))))

--

-- -- that is: /\T:*,/\a:T,/\b:T,/\(a = b),b = a

--

-- proof of f_equal, x = y -> f x = f y

--

-- > (\x:nat,\y:nat,\f:(/\n:nat,nat),\e:eq nat x y,eqElim nat x (\y:nat,eq nat (f x) (f y)) (refl nat (f x)) y e)

-- Pi (Free (Global "nat")) (Scope (

-- Pi (Free (Global "nat")) (Scope (

-- Pi (Pi (Free (Global "nat")) (Scope (Free (Global "nat")))) (Scope (

-- Pi (App (App (App (Free (Global "eq")) (Free (Global "nat"))) (Bound 2)) (Bound 1)) (Scope (

-- App (App (App (Free (Global "eq")) (Free (Global "nat"))) (App (Bound 1) (Bound 3))) (App (Bound 1) (Bound 2))))))))))

--

-- Induction proof that: n + 0 = n

--

-- > (natElim (\n:nat,eq nat (add n Z) n) (refl nat Z) (\n:nat,\e:eq nat (add n Z) n,(\x:nat,\y:nat,\f:(/\n:nat,nat),\e:eq nat x y,eqElim nat x (\y:nat,eq nat (f x) (f y)) (refl nat (f x)) y e) (add n Z) n S e))

-- Pi (Free (Global "nat")) (Scope (

-- App (App (App (Free (Global "eq")) (Free (Global "nat")))

-- (App (App (App (App (Free (Global "natElim"))

-- (Lam Type (Scope (Free (Global "nat")))))

-- (Free (Global "Z")))

-- (Lam Type (Scope (Free (Global "S")))))

-- (Bound 0)))

-- (Bound 0)))

## Sunday, 1 March 2009

### How to halve a number

(*

* How to halve a number

* ---------------------

*)

(* (* just so we start on the same page, what is a number? *)

*

* Inductive nat : Set :=

* | O : nat

* | S : nat -> nat.

*)

Module Type Halver.

Parameter half : nat -> nat.

Axiom e1 : half 0 = 0.

Axiom e2 : half 1 = 0.

Axiom e3 : forall n, half (S (S n)) = S (half n).

(* An equational specification is given as a module

* so now it is possible it give various implementations

*)

End Halver.

Module By_Default : Halver.

Fixpoint half (n : nat) : nat :=

match n with

| O => O

| S O => O

| S (S n) => S (half n)

end.

Definition e1 : half 0 = 0 := refl_equal _.

Definition e2 : half 1 = 0 := refl_equal _.

Definition e3 n : half (S (S n)) = S (half n) := refl_equal _.

(* This is the sort of thing one would normally start with,

* pattern match on a term and use recursive calls on a

* structurally smaller term, the idea of 'structurally smaller'

* is part of the type rules actually and it is the transitive

* closure of (1) p_i < C p_1 p_2 ... p_n

* and (2) f x < f

* rule (1) and (2) say that a term in smaller than any

* construction built from it, a construction being something

* literally built from constructor or a function

*)

End By_Default.

Module By_Divine_Inspiration : Halver.

(* Definition half n :=

* fst (nat_rect (fun _ => (nat * nat)%type)

* (O, O)

* (fun n Pn => let (p,q) := Pn in (q, S p))

* n).

* (* it seems impossible to prove this one correct *)

*)

Definition half n :=

fst (nat_rect (fun _ => (nat * nat)%type)

(O, O)

(fun n Pn => (snd Pn, S (fst Pn)))

n).

Definition e1 : half 0 = 0 := refl_equal _.

Definition e2 : half 1 = 0 := refl_equal _.

Definition e3 n : half (S (S n)) = S (half n) := refl_equal _.

End By_Divine_Inspiration.

Module By_Devising_Data : Halver.

(* This way seems like a 'view' in the sense of Wadler,

* The structure of the data make the program definition

* straight-forward.

*)

Inductive nat' : Set :=

| Zero : nat'

| One : nat'

| SuccSucc : nat' -> nat'.

Definition zero := Zero.

Fixpoint succ (n : nat') :=

match n with

| Zero => One

| One => SuccSucc Zero

| SuccSucc n => SuccSucc (succ n)

end.

Fixpoint nat_to_nat' (n : nat) : nat' :=

match n with

| O => zero

| S n => succ (nat_to_nat' n)

end.

Fixpoint half' (n : nat') : nat :=

match n with

| Zero => O

| One => O

| SuccSucc n => S (half' n)

end.

Definition half n := half' (nat_to_nat' n).

Definition e1 : half 0 = 0 := refl_equal _.

Definition e2 : half 1 = 0 := refl_equal _.

Lemma succ_succs n : succ (succ n) = SuccSucc n.

induction n; simpl; try reflexivity.

rewrite IHn; reflexivity.

Qed.

Theorem e3 n : half (S (S n)) = S (half n).

intro n.

unfold half.

simpl.

rewrite succ_succs.

reflexivity.

Qed.

End By_Devising_Data.

Module By_Mutual_Induction : Halver.

Inductive Even : nat -> Set :=

| EvenO : Even O

| EvenS n : Odd n -> Even (S n)

with Odd : nat -> Set :=

| OddS n : Even n -> Odd (S n).

Fixpoint even_or_odd n : Even n + Odd n :=

match n as n return Even n + Odd n with

| O => inl _ EvenO

| S n =>

match even_or_odd n with

| inl even => inr _ (OddS _ even)

| inr odd => inl _ (EvenS _ odd)

end

end.

Fixpoint halfE n (e : Even n) : nat :=

match e with

| EvenO => O

| EvenS n odd => S (halfO n odd)

end

with halfO n (o : Odd n) : nat :=

match o with

| OddS n even => halfE n even

end.

Definition half (n : nat) : nat :=

match even_or_odd n with

| inl even => halfE n even

| inr odd => halfO n odd

end.

(* Proving this program correct is quite a lot

* of work compared to the previous implementations,

* these tricky inversion lemmas which would be trivial

* to prove using pattern matching at least show some

* techniques.

*)

Lemma unique_lemma_1 (e : Even 0) : e = EvenO.

intro e.

refine (

match e as e' in Even Zero return

match Zero return Even Zero -> Prop with

| O => fun e'' : Even 0 => e'' = EvenO

| S _ => fun _ => True

end e'

with

| EvenO => refl_equal _

| EvenS _ _ => I

end

).

Qed.

Lemma unique_lemma_2 n (e : Even (S n)) : { o' : _ | e = EvenS n o' }.

intros n e.

refine (

match e as e' in Even SuccN return

match SuccN return Even SuccN -> Set with

| O => fun _ => True

| S n => fun e' : Even (S n) =>

{ o' : _ | e' = EvenS n o' }

end e'

with

| EvenO => I

| EvenS n' o' => exist _ o' (refl_equal _)

end

).

Qed.

Lemma unique_lemma_3 n (o : Odd (S n)) : { e' : _ | o = OddS n e' }.

intros n o.

refine (

match o as o' in Odd SuccN return

match SuccN return Odd SuccN -> Set with

| O => fun _ => True

| S n => fun o' : Odd (S n) =>

{ e' : _ | o' = OddS n e' }

end o'

with

| OddS n' e' => exist _ e' (refl_equal _)

end

).

Qed.

(* There is only one way to prove a number is Even

* and similarly to prove it's Odd

*)

Theorem unique_classificationE n (e e' : Even n) : e = e'

with unique_classificationO n (o o' : Odd n) : o = o'.

induction n; intros.

rewrite unique_lemma_1 with e.

rewrite unique_lemma_1 with e'.

reflexivity.

destruct (unique_lemma_2 _ e) as [o eq].

destruct (unique_lemma_2 _ e') as [o' eq'].

rewrite eq; rewrite eq'.

rewrite (unique_classificationO n o o').

reflexivity.

destruct n.

intro H; inversion H.

intros.

destruct (unique_lemma_3 _ o) as [e eq].

destruct (unique_lemma_3 _ o') as [e' eq'].

rewrite eq; rewrite eq'.

rewrite (unique_classificationE n e e').

reflexivity.

Qed.

Lemma halfE_SS n e e' : halfE (S (S n)) e' = S (halfE n e)

with halfO_SS n o o' : halfO (S (S n)) o' = S (halfO n o).

intros.

destruct (unique_lemma_2 _ e') as [o eq].

destruct (unique_lemma_3 _ o) as [e'' eq'].

subst.

simpl.

assert (e = e'') as E by apply unique_classificationE.

rewrite E; reflexivity.

intros.

destruct (unique_lemma_3 _ o') as [e eq].

destruct (unique_lemma_2 _ e) as [o'' eq'].

subst.

simpl.

assert (o = o'') as O by apply unique_classificationO.

rewrite O; reflexivity.

Qed.

Definition which P Q : P + Q -> bool :=

fun choice =>

match choice with

| inl _ => true

| inr _ => false

end.

Lemma even_or_odd_SS n : which _ _ (even_or_odd (S (S n))) = which _ _ (even_or_odd n).

induction n.

reflexivity.

simpl in *.

destruct (even_or_odd n).

reflexivity.

reflexivity.

Qed.

Definition e1 : half 0 = 0 := refl_equal _.

Definition e2 : half 1 = 0 := refl_equal _.

Theorem e3 n : half (S (S n)) = S (half n).

intro n.

unfold half.

generalize (even_or_odd_SS n).

destruct (even_or_odd n); destruct (even_or_odd (S (S n)));

simpl; intro hyp; try discriminate hyp.

apply halfE_SS.

apply halfO_SS.

Qed.

End By_Mutual_Induction.

Module By_Well_Founded_Induction : Halver.

Section Well_Founded_Relations.

Variable A : Type.

Variable R : A -> A -> Prop.

Inductive Acc (x : A) : Prop :=

| below : (forall y : A, R y x -> Acc y) -> Acc x.

(* x must be a parameter (instead of Acc : A -> Prop) for

* Set/Type elimination, but why?

*)

Definition Well_Founded := forall x : A, Acc x.

Inductive Transitive_Closure : A -> A -> Prop :=

| step x y : R x y -> Transitive_Closure x y

| transitivity x y z : R x y -> Transitive_Closure y z -> Transitive_Closure x z.

Definition Well_Founded_Induction

(W : Well_Founded)

(P : A -> Type)

(rec : forall x, (forall y, R y x -> P y) -> P x)

: forall a, P a :=

fun a => Acc_rect P (fun x _ hyp => rec x hyp) a (W a).

End Well_Founded_Relations.

Section Transitive_Relation.

Variable A : Type.

Variable R : A -> A -> Prop.

Theorem Transitive_Closure_Well_Founded :

Well_Founded A R -> Well_Founded A (Transitive_Closure A R).

Proof.

unfold Well_Founded; intros R_Acc x.

induction (R_Acc x) as [x R_ind T_ind].

apply below; intros y Trans.

induction Trans.

apply T_ind; assumption.

pose (IHTrans R_ind T_ind) as IH; inversion IH as [H']; subst.

apply H'.

apply step.

assumption.

Defined.

End Transitive_Relation.

Section Nat_Structural_Measure.

(* Looking at the definition of nat: *)

(*

* Inductive nat : Set :=

* | O : nat

* | S : nat -> nat.

*)

(* shows there is one (simple) recursive field *)

Inductive nat_step : nat -> nat -> Prop :=

| nat_step_a : forall x, nat_step x (S x).

(* the nat_step relation expresses this recursion

* every inductive type admits a step relation

* (indexed inductives define indexed relations,

* but maybe wrappig them in sigT would be better?)

*)

Theorem nat_step_Well_Founded : Well_Founded nat nat_step.

unfold Well_Founded.

induction x;

apply below; intros y Ry.

inversion Ry.

inversion Ry.

subst.

assumption.

Defined.

(* Could one generically define all Φ_step relations,

* and give a generic proof that they are all Well_Founded?

*)

(* The transitive closure of the step relation corresponds to

* the usual idea of lt/(<) on nat:

*)

Definition nat_struct := Transitive_Closure nat nat_step.

Definition nat_struct_Well_Founded :=

Transitive_Closure_Well_Founded nat nat_step nat_step_Well_Founded.

(* In general terms it is a typed version of the 'structurally smaller'

* relation in the metatheory which lets definitions like the one in

* By_Default be accepted

*)

End Nat_Structural_Measure.

(* Now the equations we want to satisy are: *)

(*

* half O := O

* half (S O) := O

* half (S (S n)) := S (half n)

*)

(* so in terms of (toned down) eliminators that is:

*

* half n := nat_case n

* O

* (fun n' Pn' =>

* nat_case n'

* O

* (fun n'' Pn'' =>

* S (half n''))).

*

* clearly that definition is not acceptable with

* no termination argument what-so-ever! so to

* rectify that, we can let P express that this

* number is smaller than n, and use well founded

* induction on the size measure.

*)

Lemma SSsmaller m : nat_struct m (S (S m)).

intro m.

apply transitivity with (S m).

constructor.

apply step.

constructor.

Defined.

Definition half := Well_Founded_Induction nat nat_struct nat_struct_Well_Founded

(fun _ => nat)

(nat_rec (fun x => (forall y : nat, nat_struct y x -> nat) -> nat)

(fun rec => O)

(fun n Pn rec =>

nat_rec (fun x => (forall y : nat, nat_struct y (S x) -> nat) -> nat)

(fun rec => O)

(fun m Pm rec => S (rec m (SSsmaller m)))

n

rec)).

(*

* The definition may look a bit complicated but I think this could

* be computed automatically from the equational specification,

* the method seems like a good way to compile dependent pattern matching

* into a basic type theory though, so that we don't have to use something

* like a 'match' construct.

*)

Definition e1 : half 0 = 0 := refl_equal _.

Definition e2 : half 1 = 0 := refl_equal _.

Definition e3 n : half (S (S n)) = S (half n) := refl_equal _.

(*

* It's not luck (at least I think it's not, and I really hope so)

* that the recursive equality is provable by reflexivity, but it

* does require the SSsmaller proof term to be canonical (made up

* of constructors), since every proof that a term is structurally

* smaller than another can be a canonical proof that poses no

* problem. I don't yet know how it will interact with dependent

* pattern match specialization though. (I can't think of any

* functions over inductive families that aren't 'step' recursive

* though, does anyone know of some examples?)

*)

End By_Well_Founded_Induction.