Composition versus Inheritance

A Comparative Look at Two Fundamental Ways to Relate Classes

by Bill Venners
October 15, 1998

First published in JavaWorld, October 1998
Summary
In this installment of my Design Techniques column, I analyze the flexibility and performance implications of inheritance and composition, and I give guidelines on the appropriate use of each.

One of the fundamental activities of any software system design is establishing relationships between classes. Two fundamental ways to relate classes are inheritance and composition. Although the compiler and Java virtual machine (JVM) will do a lot of work for you when you use inheritance, you can also get at the functionality of inheritance when you use composition. This article will compare these two approaches to relating classes and will provide guidelines on their use.

First, some background on the meaning of inheritance and composition.

About inheritance
In this article, I'll be talking about single inheritance through class extension, as in:

class Fruit {

    //...
}

class Apple extends Fruit {

    //...
}

In this simple example, class Apple is related to class Fruit by inheritance, because Apple extends Fruit. In this example, Fruit is the superclass and Apple is the subclass.

I won't be talking about multiple inheritance of interfaces through interface extension. That topic I'll save for next month's Design Techniques article, which will be focused on designing with interfaces.

Here's a UML diagram showing the inheritance relationship between Apple and Fruit:

Inheritance relationship
Figure 1. The inheritance relationship

About composition
By composition, I simply mean using instance variables that are references to other objects. For example:

class Fruit {

    //...
}

class Apple {

    private Fruit fruit = new Fruit();
    //...
}

In the example above, class Apple is related to class Fruit by composition, because Apple has an instance variable that holds a reference to a Fruit object. In this example, Apple is what I will call the front-end class and Fruit is what I will call the back-end class. In a composition relationship, the front-end class holds a reference in one of its instance variables to a back-end class.

The UML diagram showing the composition relationship has a darkened diamond, as in:

Composition relationship
Figure 2. The composition relationship

Dynamic binding, polymorphism, and change
When you establish an inheritance relationship between two classes, you get to take advantage of dynamic binding and polymorphism. Dynamic binding means the JVM will decide at runtime which method implementation to invoke based on the class of the object. Polymorphism means you can use a variable of a superclass type to hold a reference to an object whose class is the superclass or any of its subclasses.

One of the prime benefits of dynamic binding and polymorphism is that they can help make code easier to change. If you have a fragment of code that uses a variable of a superclass type, such as Fruit, you could later create a brand new subclass, such as Banana, and the old code fragment will work without change with instances of the new subclass. If Banana overrides any of Fruit's methods that are invoked by the code fragment, dynamic binding will ensure that Banana's implementation of those methods gets executed. This will be true even though class Banana didn't exist when the code fragment was written and compiled.

Thus, inheritance helps make code easier to change if the needed change involves adding a new subclass. This, however, is not the only kind of change you may need to make.

Changing the superclass interface
In an inheritance relationship, superclasses are often said to be "fragile," because one little change to a superclass can ripple out and require changes in many other places in the application's code. To be more specific, what is actually fragile about a superclass is its interface. If the superclass is well-designed, with a clean separation of interface and implementation in the object-oriented style, any changes to the superclass's implementation shouldn't ripple at all. Changes to the superclass's interface, however, can ripple out and break any code that uses the superclass or any of its subclasses. What's more, a change in the superclass interface can break the code that defines any of its subclasses.

For example, if you change the return type of a public method in class Fruit (a part of Fruit's interface), you can break the code that invokes that method on any reference of type Fruit or any subclass of Fruit. In addition, you break the code that defines any subclass of Fruit that overrides the method. Such subclasses won't compile until you go and change the return value of the overridden method to match the changed method in superclass Fruit.

Inheritance is also sometimes said to provide "weak encapsulation," because if you have code that directly uses a subclass, such as Apple, that code can be broken by changes to a superclass, such as Fruit. One of the ways to look at inheritance is that it allows subclass code to reuse superclass code. For example, if Apple doesn't override a method defined in its superclass Fruit, Apple is in a sense reusing Fruit's implementation of the method. But Apple only "weakly encapsulates" the Fruit code it is reusing, because changes to Fruit's interface can break code that directly uses Apple.

The composition alternative
Given that the inheritance relationship makes it hard to change the interface of a superclass, it is worth looking at an alternative approach provided by composition. It turns out that when your goal is code reuse, composition provides an approach that yields easier-to-change code.

Code reuse via inheritance
For an illustration of how inheritance compares to composition in the code reuse department, consider this very simple example:

class Fruit {

    // Return int number of pieces of peel that
    // resulted from the peeling activity.
    public int peel() {

        System.out.println("Peeling is appealing.");
        return 1;
    }
}

class Apple extends Fruit {
}

class Example1 {

    public static void main(String[] args) {

        Apple apple = new Apple();
        int pieces = apple.peel();
    }
}

When you run the Example1 application, it will print out "Peeling is appealing.", because Apple inherits (reuses) Fruit's implementation of peel(). If at some point in the future, however, you wish to change the return value of peel() to type Peel, you will break the code for Example1. Your change to Fruit breaks Example1's code even though Example1 uses Apple directly and never explicitly mentions Fruit.

Here's what that would look like:

class Peel {

    private int peelCount;

    public Peel(int peelCount) {
        this.peelCount = peelCount;
    }

    public int getPeelCount() {

        return peelCount;
    }
    //...
}

class Fruit {

    // Return a Peel object that
    // results from the peeling activity.
    public Peel peel() {

        System.out.println("Peeling is appealing.");
        return new Peel(1);
    }
}

// Apple still compiles and works fine
class Apple extends Fruit {
}

// This old implementation of Example1
// is broken and won't compile.
class Example1 {

    public static void main(String[] args) {

        Apple apple = new Apple();
        int pieces = apple.peel();
    }
}

Code reuse via composition
Composition provides an alternative way for Apple to reuse Fruit's implementation of peel(). Instead of extending Fruit, Apple can hold a reference to a Fruit instance and define its own peel() method that simply invokes peel() on the Fruit. Here's the code:

class Fruit {

    // Return int number of pieces of peel that
    // resulted from the peeling activity.
    public int peel() {

        System.out.println("Peeling is appealing.");
        return 1;
    }
}

class Apple {

    private Fruit fruit = new Fruit();

    public int peel() {
        return fruit.peel();
    }
}

class Example2 {

    public static void main(String[] args) {

        Apple apple = new Apple();
        int pieces = apple.peel();
    }
}

In the composition approach, the subclass becomes the "front-end class," and the superclass becomes the "back-end class." With inheritance, a subclass automatically inherits an implemenation of any non-private superclass method that it doesn't override. With composition, by contrast, the front-end class must explicitly invoke a corresponding method in the back-end class from its own implementation of the method. This explicit call is sometimes called "forwarding" or "delegating" the method invocation to the back-end object.

The composition approach to code reuse provides stronger encapsulation than inheritance, because a change to a back-end class needn't break any code that relies only on the front-end class. For example, changing the return type of Fruit's peel() method from the previous example doesn't force a change in Apple's interface and therefore needn't break Example2's code.

Here's how the changed code would look:

class Peel {

    private int peelCount;

    public Peel(int peelCount) {
        this.peelCount = peelCount;
    }

    public int getPeelCount() {

        return peelCount;
    }
    //...
}

class Fruit {

    // Return int number of pieces of peel that
    // resulted from the peeling activity.
    public Peel peel() {

        System.out.println("Peeling is appealing.");
        return new Peel(1);
    }
}

// Apple must be changed to accomodate
// the change to Fruit
class Apple {

    private Fruit fruit = new Fruit();

    public int peel() {

        Peel peel = fruit.peel();
        return peel.getPeelCount();
    }
}

// This old implementation of Example2
// still works fine.
class Example1 {

    public static void main(String[] args) {

        Apple apple = new Apple();
        int pieces = apple.peel();
    }
}

This example illustrates that the ripple effect caused by changing a back-end class stops (or at least can stop) at the front-end class. Although Apple's peel() method had to be updated to accommodate the change to Fruit, Example2 required no changes.

Comparing composition and inheritance
So how exactly do composition and inheritance compare? Here are several points of comparison:

  • It is easier to change the interface of a back-end class (composition) than a superclass (inheritance). As the previous example illustrated, a change to the interface of a back-end class necessitates a change to the front-end class implementation, but not necessarily the front-end interface. Code that depends only on the front-end interface still works, so long as the front-end interface remains the same. By contrast, a change to a superclass's interface can not only ripple down the inheritance hierarchy to subclasses, but can also ripple out to code that uses just the subclass's interface.
  • It is easier to change the interface of a front-end class (composition) than a subclass (inheritance). Just as superclasses can be fragile, subclasses can be rigid. You can't just change a subclass's interface without making sure the subclass's new interface is compatible with that of its supertypes. For example, you can't add to a subclass a method with the same signature but a different return type as a method inherited from a superclass. Composition, on the other hand, allows you to change the interface of a front-end class without affecting back-end classes.
  • Composition allows you to delay the creation of back-end objects until (and unless) they are needed, as well as changing the back-end objects dynamically throughout the lifetime of the front-end object. With inheritance, you get the image of the superclass in your subclass object image as soon as the subclass is created, and it remains part of the subclass object throughout the lifetime of the subclass.
  • It is easier to add new subclasses (inheritance) than it is to add new front-end classes (composition), because inheritance comes with polymorphism. If you have a bit of code that relies only on a superclass interface, that code can work with a new subclass without change. This is not true of composition, unless you use composition with interfaces. Used together, composition and interfaces make a very powerful design tool. I'll talk about this approach in next month's Design Techniques article.
  • The explicit method-invocation forwarding (or delegation) approach of composition will often have a performance cost as compared to inheritance's single invocation of an inherited superclass method implementation. I say "often" here because the performance really depends on many factors, including how the JVM optimizes the program as it executes it.
  • With both composition and inheritance, changing the implementation (not the interface) of any class is easy. The ripple effect of implementation changes remain inside the same class.

Choosing between composition and inheritance
So how do all these comparisons between composition and inheritance help you in your designs? Here are a few guidelines that reflect how I tend to select between composition and inheritance.

Make sure inheritance models the is-a relationship
My main guiding philosophy is that inheritance should be used only when a subclass is-a superclass. In the example above, an Apple likely is-a Fruit, so I would be inclined to use inheritance.

An important question to ask yourself when you think you have an is-a relationship is whether that is-a relationship will be constant throughout the lifetime of the application and, with luck, the lifecycle of the code. For example, you might think that an Employee is-a Person, when really Employee represents a role that a Person plays part of the time. What if the person becomes unemployed? What if the person is both an Employee and a Supervisor? Such impermanent is-a relationships should usually be modelled with composition.

Don't use inheritance just to get code reuse
If all you really want is to reuse code and there is no is-a relationship in sight, use composition.

Don't use inheritance just to get at polymorphism
If all you really want is polymorphism, but there is no natural is-a relationship, use composition with interfaces. I'll be talking about this subject next month.

Next month
In next month's Design Techniques article, I'll talk about designing with interfaces.

A request for reader participation
I encourage your comments, criticisms, suggestions, flames -- all kinds of feedback -- about the material presented in this column. If you disagree with something, or have something to add, please let me know.

You can either participate in a discussion forum devoted to this material or e-mail me directly at bv@artima.com.

Resources

This article was first published under the name Composition versus Inheritance: Which One Should You Choose? in JavaWorld, a division of Web Publishing, Inc., October 1998.

Talk back!

Have an opinion? Be the first to post a comment about this article.

About the author

Bill Venners has been writing software professionally for 12 years. Based in Silicon Valley, he provides software consulting and training services under the name Artima Software Company. Over the years he has developed software for the consumer electronics, education, semiconductor, and life insurance industries. He has programmed in many languages on many platforms: assembly language on various microprocessors, C on Unix, C++ on Windows, Java on the Web. He is author of the book: Inside the Java Virtual Machine, published by McGraw-Hill.