Intro to Swift

Intro to Swift

This workshop will be a general introduction to the Swift programming language.


This workshop assumes:

Set up instructions

Contributors: Jared Khan, Martin Hartt
View slides
View code examples on GitHub
License: MIT

Why Swift?

Swift is designed to be:

a programming language to empower everyone to turn their ideas into apps on any platform.

Swift was designed by Apple and has taken over (from Objective-C) as the go-to language for programming Apple platforms. Whilst the vast majority of Swift code is written for iOS and MacOS, its reach doesn't end there. I want to show you today that Swift is a powerful and usable general-purpose programming language that is not constrained to Apple platforms.

Swift is open source. That is to say, anybody can contribute to its development by filing issues, proposing new features and changes, and by directly contributing code. You can find the compiler, standard library and core libraries on GitHub. If you choose to, you can participate in the process of shaping the future of this young language.

If you're interested in mobile app development then Swift programming is an essential skill right now but even outside of that context Swift is a great example of a modern programming language. Learning Swift may help you understand many other languages from which it draws inspiration.

What we will cover

This workshop covers most of the content of the 'A Swift Tour' section of The Swift Programming Language. This is the go-to resource for learning the interesting bits of Swift and if you're interested in learning more then that should be your first stop.


Most of the time when you write Swift, you are using a combination of:

If you're on a Mac, Xcode is the best way of reading the documentation for these things. Open Xcode and go to Help > Developer Documentation and you can search for what you need.

If you're not on a Mac, it can be tricky to find the right docs though they are online:

In practice, Google is usually a quicker way to find the information you need.

Getting started

To follow along with this workshop, you'll first need to download the example repository. Unzip it into your machine open the folder in your terminal of choice.

Alternatively, you can clone the example repository:

$ git clone
Cloning into workshop-intro-to-git...
$ cd workshop-intro-to-git

Running Swift

We'll also need to make sure we can run Swift

REPL mode

Runs a 'read, evaluate, print' loop. Type instructions and press enter to run them. Note that this mode may not work in Linux containers (but should do on native Linux machines).

$ swift

Interpreted mode

File is read and executed line-by-line.

$ swift test.swift

Compiled mode

The swift file is compiled into a native executable.

$ swiftc test.swift -o test
$ ./test

In this workshop we'll be building a small example app that touches on many of the features of Swift that we're going to cover. The app is a directory of public toilets in Cambridge. We'll tell it where we are and a few other details and it'll give us our nearest public toilet!

In your terminal, navigate to examples/Lavy. Make sure you can run it by running:

$ swift run

This is running swiftc on a bunch of different files for us and then running the compiled binary. Handy! (this is a feature of 'Swift Package Manager' which we won't talk too much about here but you can read more here).

You should see something like the following:

Compile Swift Module 'Lavly' (5 sources)
Linking ./.build/x86_64-apple-macosx10.10/debug/Lavly
      .__   .-".
   (o\"\  |  |
      \_\ |  |
     _.---:_ |
    ("-..-" /
     "-.-" /
       /   |
       "--"  Welcome to Lavly!
Enter your latitude: 

You'll notice that some features of the app aren't implemented yet. You're going to implement them along the way. I'll mention when you know what you need to know to implement certain bits. Open the Lavly project in your favourite text editor and lets get started.

Basic syntax

Let's take a look at some of the basic syntax of Swift.

Note: After this section you will be able to complete TODO items 1-5 in the code

Unlike other languages, semicolons are optional and generally not preffered.


Values are constant variables which never change value (are 'immutable'). They are defined with the let keyword. Attempts to change the value of a constant will fail at compile time.

let name = "Richard"
name = "Hal" // error: cannot assign to value: 'name' is a 'let' constant


Variables store values which can be changed (are 'mutable'). They are defined with var.

var otherName = "Richard"
otherName = "Hal" // This succeeds

Explicitly declaring type

We can define the 'type' of a variable explicitly using a colon followed by the type name:

var title: String
title = "Mr"

Notice that we didn't have to do this before since Swift can infer the type of the variable from the type of the value that you assign to it. e.g. if we assign a String to title then it knows that title should be of type String

In this case, we have declared the variable before initialising it with a value so we do need to explicitly tell Swift what it's type is going to be.


Some other language require arrays to be of a fixed size:

int intArray[] = new integer[4];
// Fixed size, cannot expand

This doesn't happen in Swift. We can add elements to arrays on the fly and Swift will efficiently handle changing the size of the array for us.

var intArray: [Int]
intArray = [1, 2, 3]
intArray += [4, 5, 6]
// intArray == [1, 2, 3, 4, 5, 6]
// intArray[0] == 1

Control Flow

For loops

In other languages such as Java, we would use a for construct like this:

// Some other language...
for (int i = 0; i < n; i++) {
  // Use i

This is a bit clunky when our intention is to do something with each value from 0 to n. In Swift, we utilise the native ranges, which expresses loops in a more concise and expressive way.

let n = 5
for i in 0 ..< n {
// 0
// 1
// 2
// 3
// 4

Here 0 ..< n indicates the range of numbers from 0 up to (but not including) n and the for i in 0 ..< n iterates through each value in that range. We can also do an inclusive range like so: 0 ... n

Switch statements

switch statements allow us to branch our code based on the values of a variable or expression. Here we want to produce a response based on the value of a favouriteFood string.

let favouriteFood = "ratatouille"

switch favouriteFood {
  case "salad":
    print("Nobody likes a liar")
  case "candy floss":
    print("You’ve a sweet tooth")
    print("Great choice!")

Note these cases cover all possibilities and the Swift compiler knows that. In fact, it won't let us write a switch statement that doesn't cover all possibilities. Let's see what happens when we try to remove that default case:

switch favouriteFood {
  case "salad":
    print("Nobody likes a liar")
  case "candy floss":
    print("You’ve a sweet tooth")
// error: switch must be exhaustive, consider adding a default clause

We can combine ranges and switch statements to check if a value is in a certain range:

let yourAge: Int = 20 // Set this to be your age

switch yourAge {
  case 0: // This will only match the number 0 
    print("You are a baby")
  case 1 ... 3: // This will match numbers 1, 2 and 3.
    print("You are a toddler")
  case 4 ... 11:
    print("You are a child")
  case 12 ..< 18: // Here we are checking for a range which is >= 12, but < 18.
    print("You are a teenager")
  case 18...: // As no upper bound is specified, this matches all >= 18
    print("You are an adult")
  default: // For cases when there is a negative number
    print("You can't have a negative age")

Guard Statements

Whilst we're talking about control flow, I want to show you an example of where the Swift compiler makes life a lot easier for us sometimes.

guard(divisor != 0) else {
  print("Can't divide by zero")
print(5 / divisor)

Guard statements are like if statements but instead of executing the code inside the block when the condition does hold, a guard statement makes sure that the condition holds and then runs the code inside the block otherwise.

Here's the cool bit: the type system will make sure that the else block either never terminates or it gets us out of the current scope. In this case, that has been done with a return but it could have been done by throwing an error or breaking or continueing out of a current loop iteration.

Why can't we just use an if?

if(divisor == 0) {
	print("Can't divide by zero")
	// Oops!
print(5 / divisor)

Whoops! We forgot to return early when we recognised the 0. This compiles just fine but then crashes at runtime when we try to do this division. With guard on the other hand:

guard(divisor != 0) else {
  print("Can't divide by zero")
print(5 / divisor)

// error: 'guard' body may not fall through, consider using a 'return' or 'throw' to exit the scope

The compiler has caught this for us and saved us from another runtime error. Hooray!


In most other languages, any variable can potentially be empty (i.e. containing null).

// Some other language...
TrueLove richardsLove = null

Here we are representing Richard's true love in this variable richardsLove. It so happens that Richard doesn't have a true love right now so this is set to null.

When we come to use this value, we make sure to remember to check if it's null.

// Some other language...
if (richardsLove == null) {
  // Do nothing
} else {
  // sendRoses expects a TrueLove parameter

But what if we forget to make this check?

// Some other language...
TrueLove richardsLove = null
// forget to check for null
// sendRoses expects TrueLove param

The compiler will let you do this but it will catch fire at runtime! Since sendRoses expects a TrueLove parameter, this operation fails resulting in a NullPointerException. These are bugs that we might not catch before we ship our code!

In Swift, we define that a variable can be optional with a ? next to the type. Optionals can be safely used with the if let construct, where the body is run only when the variable isn't nil. This is called optional binding.

let richardsLove: TrueLove? = nil

// sendRoses expects TrueLove param
// Do an ‘optional binding’
if let recipient = richardsLove {
  sendRoses(to: recipient)

recipient is set to the value of richardsLove if it is not nil. Within the if statement, recipient has type TrueLove (i.e. non-optional)

Working with Optionals

These optional things come up a lot so it's useful to have a few ways of manipulating them.

Default Values

We can define default values with the ?? operator. In the following example, when richardsLove doesn't exist, it will assign TrueLove("Taylor Swift").

let richardsLove: TrueLove? = nil

let definiteLove = richardsLove ?? TrueLove("Taylor Swift")
// definiteLove: TrueLove (non-optional)

Force Unwrapping

If we are 100% totally sure the optional isn't nil, we can force unwrap with !.

let definiteLove = richardsLove!

Again, definiteLove here would be of non-optional type because we have force unwrapped it. Note in this case that we could not assume that richardsLove would not be nil so this example would crash at runtime.

We call this force unwrapping because it's as if the type is 'wrapped' in this optional box and we want to get it out.

Optional Chaining

What if we wanted to get the name of richardsLove? Of ocurse this is also an optional value since Richard may not have a true love. We can't just say since richardsLove could be nil and nil doesn't know how to deal with .name. This is where optional chaining comes in.

let richardsLoveName = richardsLove?.name

With the inclusion of that little question mark, this will evaluate to nil if richardsLove is nil, and the name of the TrueLove otherwise.

Now say for some reason we wanted to get the middle name of Richard's true love and then convert to lowercase. Richard might not have a true love, his true love might not have a middle name, but optional chaining has us covered:

let loverMiddleName = richardsLove?.middleName?.lowercased()

If Richard doesn't have a true love, or if he does but they don't have a middle name, then a 'link in the chain' is missing and the whole expression evaluates to nil. Otherwise we get the value we wanted.

Functions and Closures


We can define functions with the func keyword. The argument and return types are required. If a function does not return a value, don't give a return type.

func checkAnagram(string1: String, string2: String) -> Bool {
  return string1.sorted() == string2.sorted()

This function checks if two strings are anagrams of one another.

The function can then be called like so:

checkAnagram(string1: "rat", string2: "tar") // true
checkAnagram(string1: "rat", string2: "ear") // false

Exercise: You can now complete TODO items 1-5 in the Lavly Code

Having to type string1 and string2 when we use this function seems unnecessary, it would be clear from context what we mean if we called: checkAnagram("rat", "tar") but obviously within the body of the function we still need to have access to string1 and string2.

Swift lets us define both a 'parameter name' (the thing you use in the body of the function) and an 'argument label' (the label you use when calling the function).

For example we could call them left and right to the outside world and not change our body:

func checkAnagram(left string1: String, right string2: String) -> Bool {
  return string1.sorted() == string2.sorted()

And use the function like so:

checkAnagram(left: "rat", right: "tar") // true

By default the argument label is the same as the parameter name. In this case it would make sense not to have an argument label because the things we are passing in to this function are clear from context (we know an anagram is a relation between two strings). We can use the _ symbol to specify this:

func checkAnagram(_ string1: String, _ string2: String) -> Bool {
  return string1.sorted() == string2.sorted()

And use the function like so:

checkAnagram("rat", "tar") // true

Much neater!

Exercise: You can now complete TODO item 6 in the Lavly Code


Closures are like functions but without a name. Let's define a closure that takes an integer and returns whether it is even or not:

let isEven = { (number: Int) -> Bool in
	return number % 2 == 0

This can then be called like so:

isEven(3) // false

Just the same as normal function syntax! So why would you use closures?

Closures really come in to their own when you have a function that you're only going to need once. Arrays in Swift have a filter method which take a closure as an argument. This closure has to return a Bool. The method runs this function on each element in the array and gives us back only the elements for which the function returned true:

[1, 8, 5, 3, 3, 6, 7].filter({ (number: Int) -> Bool in
	return number % 2 == 0
// [8, 6]

Swift lets us write this in a much more concise way.

Swift knows that this is a filter on an array of Ints so we don't need to explicitly specify that it takes Ints or returns Bool:

[1, 8, 5, 3, 3, 6, 7].filter({ number in
	return number % 2 == 0

Since the body of our closure is only one line, Swift lets us omit the return statement:

[1, 8, 5, 3, 3, 6, 7].filter({ number in
	number % 2 == 0

Here we've given the parameter to our closure the name number but if we don't give parameters names, Swift gives them default names of $0, $1 etc. in order:

[1, 8, 5, 3, 3, 6, 7].filter({ $0 % 2 == 0 })

Finally, when the last parameter to a function is a closure, Swift lets us break it out of the round brackets:

[1, 8, 5, 3, 3, 6, 7].filter { $0 % 2 == 0 }

Neat! Every step along the way here compiles and gives the same result as our original filter code.

Another example

The map method on Array runs the given closure on each element of an Array and gives us an Array of the results. Here we use it to double each element:

[3, 4, 7].map { $0 * 2 }
// [6, 8, 14]

Exercise: You can now complete TODO items 7-10 in the Lavly Code.

Note: The remainder of these notes cover some other interesting bits of Swift to help you understand the rest of the code in the project

Classes, Structures, Enumerations


A class is like a template for an object. We define one with the class keyword. Classes can contain properties (values, variables) and methods (functions). init is a special method which defines the constructor.

class Student {
  let name: String
  var triposPart: String

  func read() {
    print("I'm reading and definitely not watching Netflix")

  init(name: String, triposPart: String) { = name
    self.triposPart = triposPart

We can then use this class and its initialiser like so:

let richard = Student(name: "Richard", triposPart: "1B")
// "I'm reading and definitely not watching Netflix"


Structs are similar to classes. We can define properties, methods, and initialisers.

struct Instructor {
  let name: String
  var module: String

  init(name: String, lecturingIn module: String) { = name
    self.module = module

  func teach() {

let hal = Instructor(name: "Hal", lecturingIn: "Swift")

If we do not provide an initialiser, we get a default initialiser for free that simply takes values for all the stored properties and assigns them:

struct Instructor {
  let name: String
  var module: String
  func teach() {

let hal = Instructor(name: "Hal", module: "Swift")

Wait... so what's the difference between Structs and Classes?

Exercise: Look at the implementation of the Toilet model in Toilet.swift


Imagine we are writing a poker app and we need to encode the set of possible card suits. In another language we might do something like this:

let SUIT_SPADES   = 0
let SUIT_CLUBS    = 1
let SUIT_HEARTS   = 2

let aceOfSpadesSuit = SUIT_SPADES
// Yucky!

Here we've assigned integers to each possible suit so that each has a unique value. The trouble here is that the type of aceOfSpacesSuit, as far as the compiler is concerned, is an integer. We might might have some code somewhere that relies on suits being between 0 and 3 but someone could easily accidentally set queenOfHeartsSuit to 4 and then we'll be in trouble!

In Swift we can define this in a way that is easier to read for us, and that tells the compiler what's going on:

enum CardSuit {
  case spades
  case clubs
  case hearts
  case diamonds

let aceOfSpadesSuit = CardSuit.spades

These aren't integers or strings, they are a new type of value that we have created to represent card suits. Neat!

Enums are more powerful than they first let on. We can add methods and initialisers if we want to! Say we want to define a nicely formatted description of each suit. Let's define a function:

enum CardSuit {
  case spades, clubs, hearts, diamonds

  func suitName() -> String {
	switch self {
	case CardSuit.spades:
		return "Spades"
	case CardSuit.clubs:
		return "Clubs"
	case CardSuit.hearts:
		return "Hearts"
		return "Diamonds"

self here refers to this specific instance of CardSuit.

Note that we haven't added a default case to this switch like we have done in the past but this compiles just fine because the compilers knows that these are the only cases of CardSuit.

The Swift compiler is smart enough to infer a type for the enums so actually we don't need to put CardSuit in every case:

enum CardSuit {
  case spades, clubs, hearts, diamonds

  func suitName() -> String {
    switch self {
      case .spades: // <- Note didn't write 'CardSuit'
        return "Spades"
      case .clubs:
        return "Clubs"
      case .hearts:
        return "Hearts"
      case .diamonds:
        return "Diamonds"

Enums also let us associate values with cases. This is an example from a food menu app. Prices may be known and have a value but may also be unconfirmed (e.g. for a soup of the day). We can model this using an enum:

enum Price {
  case value(Int)
  case unconfirmed

  func description() -> String {
    switch self {
      // The `let x` here gets the value out for us so we can use it
      case .value(let x):
        return "Price is \(x) pence"
      case .unconfirmed:
        return "See the blackboard for the price"

let caviarPrice = Price.value(2300)
let soupOfTheDayPrice = Price.unconfirmed

Note: Does this look like anything we've seen before? It turns out Optional is actually implemented behind the scenes as an enum.


Extensions allow us to add additional methods to existing classes, structs and enums.

extension String {
  func isPalindrome() -> Bool {
    return self == String(self.characters.reversed())

// "hello".isPalindrome() == false
// "racecar".isPalindrome() == true


Protocols let us define a set of requirements that other types can declare conformance too. For example, the standard library includes an Equatable protocol:

public protocol Equatable {
  static func == (lhs: Self, rhs: Self) -> Bool

This protocol has only one requirement, an equality function ==. The built in Int type conforms to this protocol and that's why we can use things like this:

[1, 3, 4].contains(3) // true

Under the hood, the contains function is running the equality function on each element until it finds a match. It wouldn't be able to do this if Int wasn't Equatable.

We can make our own types conform to Equatable simply by providing that one static method and declaring conformance with ": Equatable":

struct Instructor: Equatable {
  let name: String
  var module: String
  static func == (lhs: Instructor, rhs: Instructor) -> Bool {
  	return == && lhs.module == rhs.module

And now we can compare two Instructors with ==

Protocols are more powerful than they first appear:

With all these features, Protocol-Oriented Programming is a real alternative (or accompaniment) to Obect-Oriented Programming that can make for simpler code.

Exercise: Check out the further reading links for things that might interest you. What other features would you want to add to this app?

Further reading