Jan 25, 2019

Basic Types

• Bool - True/False
• Char - Unicode character
• Int - 64 bit signed integer
• Integer - arbitrary-precision integer
• Float - 32-bit IEEE single-precision floating point number
• Double - 64-bit IEEE double-precision floating point number
• Tuple - finite (i.e., of a given arity) sequence of different types

Note, all types have capitalized names!

Note: :t can be used to ask for the type of any expression

-- try out
:t True
:t False || True
:t 'a'
:t 5
:t sqrt 5

Function Types

A function is a mapping (->) from one type (the domain) to another type (the range).

-- e.g.,
not  :: Bool -> Bool
even :: Int  -> Bool

A function of multiple arguments can be implemented in one of two ways:

1. A function that takes a tuple of the requisite types
foo :: (Int, Bool, Char) -> Int
1. A curried function
foo :: Int -> Bool -> Char -> Int

Note: (->) associates right-to-left, so, equivalent to:

foo :: Int -> (Bool -> (Char -> Int))

I.e., foo is a function which takes an Int and returns a function which takes a Bool …

Function application (space) associates left-to-right, so:

foo 5 True 'a'

is equivalent to:

(((foo 5) True) 'a')

Inspect the types of id, const, fst, and snd. Discuss.

Polymorhic Functions

In the type a -> b, a and b are type variables. Since they are unqualified, they can be replaced with any type! A function whose type declaration contains a type variable is a polymorphic function.

Since an unqualified type variable says nothing about its actual type, you can’t do much with it (why?)

But this means the type of a polymorphic function is usually very helpful in determining what it does!

e.g., what do id, const, fst, snd do?

Inspect the types of ==, +, show, read

Overloaded Types and Type Classes

(+) :: Num a => a -> a -> a

Above, Num a is a class constraint, implying that the actual type of type variable a must belong to the type class Num.

A type that contains a class constraint is called overloaded.

A class is a collection of types that support a set of overloaded functions called methods. Each type belonging to a given class is called an instance of that class.

:i gives us information on both types and classes.

-- try out
:i Int
:i Integral
:i Num
:i Eq
:i Ord
:i Show

Function Type Declarations

Though Haskell can infer types for us, it is good practice to always explitly declare types for all our functions.

-- declare the types for the following:

nand a b = not (a && b)

sum x y = x + y

fst' (x,y) = x

disc (a,b,c) = b^2-4*a*c

quad_roots a b c = ((-b+(sqrt d))/(2*a), (-b-(sqrt d))/(2*a))
where d = disc (a,b,c)

comp a b c d = if a == b then c else d

sum_or_diff a b = if a > b then a - b else a + b