Differences

This shows you the differences between two versions of the page.

--- pp:2023:haskell:l07 [2023/04/15 15:11]
tpruteanu
+++ pp:2023:haskell:l07 [2023/04/27 01:17] (current)
mihai.udubasa [Lambda calculus as a programming language (optional)] fix typo
@@ Line 3: / Line 3: @@
 ===== 7.1 Lambda Calculus =====
-**Short RECAP:** \\
 Lambda Calculus represent a axiomatic system that can be used for very basic proofs. \\
 Given a set of variables **VARS**, a expression under lambda calculus can be: \\
@@ Line 11: / Line 10: @@
 To evaluate $\lambda$-expressions, there are two types of reduction operations:
   * **$\alpha$-conversion**: given a expression: $ \lambda x.e $, you can rename all occurences of //**x**// in //**e**// with //**y**// (used for avoiding **name collisions**).
-  * **$\beta$-reduction**: given a expression: $ \lambda x.body \ param$, you can replace all occurences of //**x**// in //**body**// with //**param**//.
+  * **$\beta$-reduction**: given a expression: $ \lambda x.body \ param$, you can replace all occurences of //**x**// in //**body**// with //**param**// (We will denote this action with: $ body[x \ / \ param] $).
-==== Booleans ====
+<note warning>
-We can encode boolean values **TRUE** and **FALSE** in lambda calculus as functions that take 2 values, **x** and **y**, and return the first (for **TRUE**) or second (for **FALSE**) value. \\
+Be careful about the difference between $ E_1 = \lambda x.e_1 \ e_2 $ **and** $ E_2 = e_1[x \ / \ e_2] $. The former denotes a expression made from a // application // between a // function // and a // expression //, while the latter is the // expression // obtained applying $ \beta $**-reduction** to the former. We say $ E_1 $ is reducible to $ E_2 $ ($ E_1 => E_2 $).
+</note>
-$ TRUE = \lambda x.\lambda y.y$ \\
+<note important>
-$ FALSE = \lambda x.\lambda y.x$ \\
+We say two // expressions // are equal, if it is possible to get one of them from the other using only **$\alpha$-conversion**.
-<note>As we defined it, **TRUE** is sometimes called the **K**-Combinator (or //Kestrel//), and **FALSE** the **KI**-Combinator (or //Kite//). </note>
+If a // expression // cannot be reduced further using $ \beta $**-reductions**, we say the expression is in $ \beta $**-normal form**.
+</note>
-<hidden>
+==== Free and bound variables ====
-Some common operation on booleans (that were discussed during the lecture) are: \\
+Take the following Scala snippet as an example:
+<code scala>
+def f(x: Int) = x + y
+</code>
+We can say that the second occurence of $ x $ is // bounded // by the $ x $ that appears as a function parameter. When we call the function, the occurence of $ x $ is replaced by the argument that was provided to $ f $. In contrast, $ y $ is a // free // variable. \\
+This code might look weird, where does $ y $ come from? What does it do? Why would we use a variable that we don't instantiate (i.e. is not bound to anything)? Well, the snippet actually comes from a broader context:
+<code scala>
+def g(x: Int, y: Int) = {
+  def f(x: Int) = x + y
+  f(x * y)
+}
+</code>
+In this new snippet we can see that all variables are //bounded//, and the free variable from before is // bounded // by the outer function, but only the // free // variable, notice that $ x $ is still bounded by the inner function, and the $ x $ parameter of $ g $ is ignored inside $ f $. \\
+The importance of // free // variables is that only // free occurences // of a sub-expression can be bounded by the outer expression. \\
 \\
-$ AND = \lambda x.\lambda y.((x \ y) \ x) $ \\
+Translating to lambda calculus, when reducing $ \lambda x.e_1 \ e_2 $ to $ e_1[x \ / \ e_2] $, only // free // occurences of $ x $ in $ e_1 $ will be replaced by $ e_2 $.
-$ OR = \lambda x.\lambda y.((x \ x) \ y) $ \\
-$ NOT = \lambda x.((x \ FALSE) \ TRUE) $ \\
+More generally, we say that:
+  * if all occurences of a variable in a expressions are // bounded //, the variable is said to be // bounded //
+  * if one occurence of a variable in a expressions is // free //, the variable is said to be // free //
-<note>
-**NOT** can also be written as: \\
 \\
-$ NOT = \lambda x.\lambda a.\lambda b.((x \ b) \ a) $ \\
+**Exercise** \\
-\\
-You can convince yourself that this works by evaluating **NOT TRUE** and **NOT FALSE**. This way of writting **NOT** is also called the **C**-Combinator (or //Cardinal//).
+**7.1.1. ** For every variable occurence, mention if it's a // free // or a // bounded // occurence:
+  - $ \lambda y.(\lambda x.x \ (x \ y)) $
+  - $ \lambda x.(x \ \lambda y.(x \ y \ z)) \ (x \ \lambda y.x) $
+  - $ \lambda f.((\lambda x.(f \ (x \ x))) \ (\lambda x.(f \ (x \ x)))) $
+==== Reduction rules ====
+Using what we learned from // free // and // bounded // variables, we can define a algorithm for $\beta$**-reduction**, given a expression $ e_1[x \ / \ e_2] $:
+^ $ e_1 $ ^ $ e_1[x \ / \ e_2] $ ^ // condition // ^
+| $ x $ | $ e_2 $ | |
+| $ y $ | $ y $ | $ x \neq y $ |
+| $ E_1 \ E_2 $ | $ E_1[x \ / \ e_2] \ E_2[x \ / \ e_2] $ | |
+| $ \lambda x.e $ | $ \lambda x.e $ | |
+| $ \lambda y.e $ | $ \lambda y.e[x \ / \ e_2] $ | $ x \neq y $, $ y $ does not appear // free // in $ e_2 $|
+| $ \lambda y.e $ | $ \{\lambda z.e[y \ / \ z]\}[x \ / \ e_2] $ | $ x \neq y $, appears // free // in $ e_2 $| ( $ z $ is a new variable that is not free in $ e $ or $ e_2 $ ) |
+==== Evaluation order ====
+**Q:** If we have multiple **redexes** in a expression, which one do we evaluate?
+**A:** We can evaluate any of them, and it is guaranteed by [[https://en.wikipedia.org/wiki/Church%E2%80%93Rosser_theorem | Church-Rosser theorem]] that if the expression is reducible, we will eventually get the same $ \beta $**-normal form**.
+To not just randomly choose **redexes**, there exist //reduction strategies//, from which we will use the **Normal Order** and **Applicative Order**: \\
+  * **Normal Order** evaluation consist of always reducing the //leftmost//, //outermost// **redex** (whenever possible, subsitute the arguments into the function body) \\
+  * **Applicative Order** evaluation consist of always reducing the //leftmost//, //innermost// **redex** (always reduce the function argument before the function itself) \\
+<note important>
+A expression of the form $ \lambda x.e_1 \ e_2 $ is also called a **redex** (reducible expression)
 </note>
-</hidden>
+**Exercise** \\
-==== Natural Numbers ====
+**7.1.2. ** Evaluate in both **Normal Order** and **Applicative Order** the following expressions:
+  - $ \lambda x.\lambda y.(x \ y \ x) \ \lambda x.\lambda y.x \ (\lambda x.\lambda y.\lambda z.(x \ z \ y) \ \lambda x.\lambda y.y)$
+  - $ \lambda x.y \ (\lambda x.(x \ x) \ \lambda x.(x \ x))$
-Church numerals represent natural numbers as **higher-order functions**. Under this representation, the number //**n**// is a function that maps **f** to its **n-fold composition**. \\
+==== Lambda calculus as a programming language (optional) ====
-\\
-$ N0 = \lambda f.\lambda x. x $ \\
-$ N1 = \lambda f.\lambda x. (f \ x) $ \\
-$ N2 = \lambda f.\lambda x. (f \ (f \ x)) $ \\
-...
-<note> Does **N0** look familiar? It's the same as **FALSE** if you rename the variables (using $\alpha$-reduction). </note>
+The [[https://en.wikipedia.org/wiki/Church%E2%80%93Turing_thesis | Church-Turing thesis]] asserts that any //computable// function can be computed using lambda calculus (or Turing Machines or equivalent models). \\
+For the curious, a series of additional exercises covering this topic can be found here: [[pp:2023:haskell:l07-extra|Lambda Calculus as a programming language]]. \\
-You can also define operation on church numerals, some (that were discussed during the lecture) are: \\
+===== 7.2 Intro to Haskell =====
-\\
-$ SUCC = \lambda n.\lambda f.\lambda x.(f \ ((n \ f) \ x)) $ \\
-$ ISZERO = \lambda n.((n \lambda x.FALSE) \ TRUE) $ \\
-$ ADD = \lambda n.\lambda m.\lambda f.\lambda x.((n \ f) ((m \ f) \ x)) $ \\
-\\
-**7.1.1** Define multiplication under church numerals: $ MULT = \lambda n.\lambda m. \ ... $ (without using the **Y**-combinator)
-**7.1.2** Define exponentiation under church numerals: $ EXP = \lambda n.\lambda m. \ ... $
+**Prequisites**: having a working haskell environment ([[pp:haskell-environment|Haskell Environment]])
-**7.1.3** (*) Define the predecessor operator, that takes a number and returns the number prior to it.
+**Haskell** is a general-purpose, purely functional programming language, that we will use for the rest of the semester to showcase functional patterns and programming styles. \\
-What's the predecessor of 0? Evaluate $ PRED \ N0 $.
+{{ :pp:2023:haskell:haskell.png?nolink&200 |}}
+This section is designed for us to get comfortable with haskell syntax, we will use several concept that we learned in Scala, such as tail-recursion, folds and maps, but this time in a purely functional context.
+==== A trip through time ====
+Remember: [[pp:2023:scala:l01|Lab 1. Introduction to Scala]]
+**7.2.1.** Implement a tail-recursive function that computes the factorial of a natural number.
+<code haskell>
+fact :: Int -> Int
+fact = undefined
+</code>
+**7.2.2.** Implement a tail-recursive function that computes the greatest common divisor of two natural numbers.
+<code haskell>
+mygcd :: Int -> Int -> Int
+mygcd a b = undefined
+</code>
+**7.2.3.** Implement the function ''mySqrt''  which computes the square root of an integer $ a $.
+==== Lists ====
+The following Scala syntax for working with lists, can be translated to Haskell as follows:
+^ Scala ^ Haskell cases ^ Haskell pattern matching ^ Haskell guards ^
+|<code scala>
+def f(l: List[Int]) = l match {
+  case Nil => ...
+  case (x::xs) => ...
+}
+</code>|<code haskell>
+f l = case l of
+  [] -> ...
+  (x:xs) -> ...
+</code> | <code haskell>
+f [] = ...
+f (x:xs) = ...
+</code> | <code haskell>
+f l | l == [] = ...
+    | otherwise = ...
+</code> |
+**7.2.4.** Implement funtions ''mymin'' and ''mymax'' that take a list of ints, and return the smallest/biggest value in the list.
+**7.2.5.** Implement a function ''unique'' that takes a list of ints, and removes all duplicates.
+**7.2.6.** Given a list of ints, return a list of strings where for each element, return:
+  * **'Fizz'** if the number is divisible by 3
+  * **'Buzz'** if the number is divisible by 5
+  * **'FizzBuzz'** if the number is divisible by 3 **and** 5
+  * a string representation of the number otherwise
+**7.2.7.** Extend the function from **7.2.6.** with the following rules:
+  * **'Bazz'** if the number is divisible by 7
+  * **'FizzBazz'** if the number is divisible by 21
+  * **'BuzzBazz'** if the number is divisible by 35
+  * **'FizzBuzzBazz'** if the number is divisible by 105
-<hidden>
-$ PRED = \lambda n.\lambda f.\lambda x.(((n \ (\lambda g.\lambda h.h \ (g \ f))) \ (\lambda u.x)) \ (\lambda v.v)) $ \\
 \\
-$ \phi = \lambda p.PAIR \ (SND \ p) \ (SUCC \ (SND \ p)) $ \\
+<hidden>
-$ PRED = \lambda.n.FST \ (n \ \phi \ (PAIR \ N0 \ N0)) $ \\
+You can test **7.2.6.** and **7.2.7.** with the following snippet, if your function is $ f $:
+<code haskell>
+f [1..n]
+</code>
 </hidden>
-\\
+<note>
+In Haskell, the list data type is denote by the type the list holds surrounded by square paranthesis.
+<code haskell>
+[Int] -- list of ints
+[Double] -- list of doubles
+[[Int]] -- list of lists of ints (matrices)
-**7.1.4** Define substraction under church numerals: $ SUB = \lambda n.\lambda m. \ ... $ (**Hint**: use $ PRED $).
+[] -- !!! not a data type, represents the empty list (Nil in Scala)
+</code>
+</note>
-What happens if you try to substract a bigger number from a smaller one? Evaluate $ SUB \ N1 \ N2 $.
+==== Types in Haskell ====
-**7.1.5** Define $ LEQ $ (less or equal). $ LEQ n m $ should return **TRUE** if $ n \leq m $ and **FALSE** if $ n > m $.
+In Haskell, functions are curried by default, **i.e.** a function:
+<code haskell>
+f a b = ...
+</code>
+is the same as:
+<code haskell>
+f = \a -> \b -> ...
+</code>
-**7.1.6** Define $ EQ $ (equality). $ EQ n m $ should return **TRUE** if $ n = m $ and **FALSE** otherwise.
+So, if $ a $ is a ''Int'' and $ b $ a ''Double'', and $ f $ returns a ''Char'', it would have the following type:
+<code haskell>
+f :: Int -> Double -> Char
+</code>
-===== 7.2 Recursion and the Y-Combinator =====
+**7.2.8.** Check the type signature of the following functions:
+  * ''foldl''
+  * ''foldr''
+  * ''filter''
+  * ''map''
-In lambda calculus, recursion is achieved using the fixed-point combinator (or **Y** combinator). \\
+<note important>
-A fixed-point combinator is a **higher-order** function that returns some fixed point of it's argument function (**x** is a fixed pointed for a function **f** if $ f(x) = x $). That means: $ f \ (fix \ f) = fix \ f $ \\
+If a function is not ambigous, ''ghc'' can infer the type signature, for **educational** purposes, going forward you will have to write signatures for all functions you define, this is considered good practice and helps prevent bugs.
-And by repeated application: $ fix \ f = f \ (f \ (... f \ (fix \ f)...)) $ \\
+</note>
-Our combinator in lambda calculus looks like this: \\
-\\
+<note tip>
-$ FIX = \lambda f.(\lambda x.f \ (x \ x)) (\lambda x.f \ (x \ x)) $
+In ''ghci'', you can check the type of a expression with: '':t''
+</note>
+===== 7.3 Brain Twisters =====
+**7.3.1.** Implement ''map'' using ''foldl'' and ''foldr''.
+<code haskell>
+mymapl :: (a -> b) -> [a] -> [b]
+mymapr :: (a -> b) -> [a] -> [b]
+</code>
+**7.3.2.** Implement ''filter'' using ''foldl'' and ''foldr''.
+<code haskell>
+myfilterl :: (a -> Bool) -> [a] -> [a]
+myfilterr :: (a -> Bool) -> [a] -> [a]
+</code>
-**7.2.1** Using the **Y**-Combinator, define a function that computes the factorial of a number **n**.
+**7.3.3.** Implement ''foldl'' using ''foldr''.
+<code haskell>
+myfoldl :: (a -> b -> a) -> a -> [b] -> a
+</code>
-**7.2.2** Using the **Y**-Combinator, define a function $ FIB $ that computes the **n**-th fibonacci number.
+**7.3.4.** Implement ''bubbleSort''.
+<code haskell>
+bubbleSort :: [Int] -> [Int]
+</code>
-===== 7.3 Intro to Haskell =====
+**7.3.5.** Implement ''quickSort''.
+<code haskell>
+quickSort :: [Int] -> [Int]
+</code>