Inheritance & Polymorphism

Inheritance in Practice

So how do we define a child class so that it inherits from a parent class?

We use the keyword extends like this:

class Shape {
    // shape class members
}

class Triangle extends Shape {
    // additional Triangle class members
}

Now Triangle has inherited traits from Shape, meaning it copied over class members from Shape. When we use inheritance to extend a subclass from a superclass, we create an "is-a" relationship from the subclass to the superclass. For example, an object of Triangle is a member of the Shape class; however an object of Shape is not necessarily an object of Triangle.

Until now, we've only covered working with one class and one file. However, most Java programs utilize multiple classes, each of which requires its own file. Only one file needs a main() method - this is the file we run.

Note: the various classes in our Java package - even though they are in different files - will have access to each other, so we can instantiate one class inside another.

Inheriting the Constructor

If the class inherits its paren'ts fields and methods, does it also inherit the constructor? Let's take a look at how the super() constructor works!

Let's say Shape has a numSides field that is set by passing an integer into the constructor. If we're instantiating a Triangle, we would want that number to always be 3, so we'd want to modify the constructor to automatically assign numSides with a value of 3.

Can we do that?

As it happens, Java has a trick up its sleeve just for this occasion: using the super() method, which acts like the parent constructor inside the class constructor:

class Triangle extends shape{

    Triangle(){
        super(3);
    }

    // additional Triangle class members
}

By passing 3 to super(), we are making it possible to instantiate a Triangle without passing in a value for numSides.

Meanwhile, super(3) (behaving as Shape(3)) will shoulder the responsibility of setting numSides to 3 for our Triangle object. It's like we called Shape(3).

It is also possible to write a constructor without making a call to any super() constructor:

class Triangle extends Shape{

    Triangle(){
        this.numSides = 3;
    }

    // additional Triangle class methods
}

In this situation, Java secretly calls the parent class' no-argument constructor (super()). So in this specific example, the Triangle() constructor first calls the Shape() constructor. That Shape() takes care of whatever business it needs to take care of. And then after that is complete, we go in and set this.numSides to 3.

If you're writing a constructor of a child class, and don't explicitly make a call to a constructor to a parent class using super, it's important to remember that Java will automatically (and secretly) call super() as the first line of your child class constructor.

Parent Class Aspect Modifiers

You may recall that Java class members use private and public access modifiers to determine whether they can be accessed from outside the class. So does a child class inherit its parent's private members?

Well, no. But there is another access modifier we can use to keep a parent class member accessible to its child classes and to files in the package it's contained in - and otherwise private: protected.

Modifier	Class	Package	Child Class	Global
`public`
`protected`
no modifier
`private`

Here's what protected looks like in use:

class Shape{

    protected double perimeter;
}

// any child class of Shape can access perimeter

In addition to access modifiers, there's another way to establish how child classes can interact with inherited parent class members: using the final keyword. If we add final before a parent class method's access modifier, we disallow any child classes from changing that method. This is helpful in limiting bugs that might occur from modifying a particular method.

Introducing Polymorphism

In Java, if Orange is a Fruit through inheritance, you can use Orange in the same contexts as Fruit like this:

String makeJuice(Fruit fruit){

    return "Apple juice and " + fruit.squeeze();

}

// inside main()
Orange orange = new Orange();
System.out.println(juicer.makeJuice(orange));

Wait, how does that work?

This is because Java incorporates the object-oriented programming principle of polymorphism. Polymorphism, which derives from Greek meaning "many forms", allows a child class to share the information and behavior of its parent class while also incorporating its own functionality.

The main advantages of polymorphic programming:

Simplifying syntax
Reducing cognitive overload for devlopers

These benefits are particularly helpful when we want to develop our own Java packages for other developers to import and use.

For example, the built-in operator + can be used for both doubles and ints. To the computer, the + means something like addDouble() for the other, but the creators of Java (and other languages) didn't want to burden us developers with recalling each individual method.

Note that the reverse situation is not true; you cannot use generic parent class inheritance where a child class instance is required. So an Orange can be used as a Fruit, but a Fruit cannot be used as an Orange.

Method Overriding

One common use of polymorphism with Java classes is something we mentioned earlier - Overriding parent class methods in a child class. Like the + operator, we can give a single method slightly different meanings for different classes. This is useful when we want our child method to have the same name as a parent class method but behave a bit differently in some way.

Let's say we have a BankAccount class that allows us to print the current balance. We want to build a CheckingAccount class that inherits the functionality of a BankAccount but with a modified printBalance() method. We can do the following:

class BankAccount{
    protected double balance;

    public BankAccount(double balanceIn){
        balance = balanceIn;
    }

    public void printBalance() {
        System.out.println("Your account balance is $" + balance);
    }
}

class CheckingAccount extends BankAccount{

    public CheckingAccount(double balance){
        super(balance);
    }

    @Override
    public void printBalance(){
        System.out.println("Your checking account balance is $" + balance);
    }
}

Notice that in order to properly override printBalance(), in CheckingAccount the method has the following in common with the corresponding method BankAccount:

Method name
Return type
Number and type of parameters

You may also have noticed the @Override keyword above printBalance() in CheckingAccount. This annotation informs the compiler that we want to override a method in the parent class. If the method doesn't exist in the parent class, we'll get a helpful error when we compile the program.

Using a Child Class as its Parent Class

An important facet of polymorphism is the ability to use a child class object where an object of its parent class is expected.

One way to do this explicitly is to instantiate a child class object as a member of the parent class. We can instantiate a CheckingAccount object as a BankAccount object like this:

BankAccount nicksAccount = new CheckingAccount(5000.00);

We can use nicksAccount as if we were an instance of BankAccount, in any situation where a BankAccount object is expected. (This would be true even if nicksAccount was instantiated as a CheckingAccount, but using the explicit child as parent syntax is most helpful when we want to declare objects in bulk.)

It is imporant to note that the compiler considers nicksAccount to be any old BankAccount. But because method overriding is handled at runtime, if we call printBalance(), we'll see something CheckingAccount specific:

Your checking account balance is $5000.00

This is because at runtime, nicksAccount is recognized as the CheckingAccount it is. So, what if CheckingAccount has a method transferToSavings() that BankAccount does not have? Can nicksAccount still use that method?

Well, no. The compiler believes that nicksAccount is just a BankAccount that doesn't have some fancy child class transferToSavings() method, so it would return an error.

Child Classes in Arrays and `ArrayList`s

Usually, when we create an array or an ArrayList, the list items all need to be the same type. But polymorphism puts a new spin on what is considered the same type...

In fact, we can put instances of different classes that share a parent class together in an array or ArrayList! For example, let's say we have a Monster parent class with a few child classes: Vampire, Werewolf, and Zombie. We can set up an array with instances of each:

Monster dracula, wolfman, zombie1;

dracula = new Vampire();
wolfman = new Werewolf();
zombie1 = new Zombie();

Monster[] monsters = {dracula, wolfman, zombie1};

We can even iterate through the list of items - regardless of subclass - and perform the same action with each item:

for (Monster monster : monsters) {

    monster.attack();
}

In the code above, we are able to call attack() on each monster in monsters despite the fact that, in the for-each loop, monster is declared as the parent class type Monster.

Child Classes in Method Parameters

When we call a method that contains parameters, the arguments we place in our method call must match the parameter type. Polymorphism gives us a little more flexibility with the arguements we can use.

If we use a superclass reference as a method parameter, we can call the method using subclass reference arguments!

For example, imagine the class ScaryStory, whose constructor takes in a reference to the Monster class:

class ScaryStory {
    Monster monster;
    String setting;

    public ScaryStory(Monster antagonist, String place){
        monster = antagonist;
        setting = place;
    }

    public static void tellStory(){
        System.out.println("Once upon a time, " + monster.name + " was at " + setting + " looking to scare some mortals.");
    }

    public static void main(String[] args) {
        Monster dracula;
        dracula = new Vampire("Dracula");
        ScaryStory countDracula = new ScaryStory(dracula, "Dracula Castle");
        countDracula.tellStory();
    }
}

In the main() method, we used a reference of the class Vampire as our argument even though the constructor requested an object of class Monster. This is allowed because Vampire is a subclass of the Monster class.

Review of Inheritance and Polymorphism

A Java class can inherit fields and methods from another class.
Each Java class requires its own file, but only one class in a Java package needs a main() method.
Child classes inherit the parent constructor by default, but it’s possible to modify the constructor using super() or override it completely.
You can use protected and final to control child class access to parent class members.
Java’s OOP principle of polymorphism means you can use a child class object like a member of its parent class, but also give it its own traits.
You can override parent class methods in the child class, ideally using the @Override keyword.
It’s possible to use objects of different classes that share a parent class together in an array or ArrayList.