OOP Concepts for Beginners: What is Composition?

Composition is one of the fundamental concepts in object-oriented programming. It describes a class that references one or more objects of other classes in instance variables. This allows you to model a has-a association between objects.

You can find such relationships quite regularly in the real world. A car, for example, has an engine and modern coffee machines often have an integrated grinder and a brewing unit.

Main benefits of composition

Given its broad use in the real world, it’s no surprise that composition is also commonly used in carefully designed software components. When you use this concept, you can:

1. reuse existing code
2. design clean APIs
3. change the implementation of a class used in a composition without adapting any external clients

Reuse existing code

The main reason to use composition is that it allows you to reuse code without modeling an is-a association as you do by using inheritance. That allows stronger encapsulation and makes your code easier to maintain as Joshua Bloch explains in the 3rd edition of his book Effective Java.

The concept of composition is often used in the real world, and it should be the same in software development. A car is not an engine; it has one. And a coffee machine has a grinder and a brewing unit, but it is none of them. The car and the coffee machine integrate an engine, grinder and brewing unit via their external APIs to compose a higher level of abstraction and provide more significant value to their users.

You can do the same in software development when you design a class to keep a reference to an object and to use it in one or more of its methods.

Design a clean API

This also enables you to design clean and easy-to-use APIs. When you compose a class, you can decide if the referenced classes become part of the API or if you want to hide them.

As I explained in my article about encapsulation, Java supports different access modifiers. It’s a common best practice to use the private modifier for all attributes, including the ones that reference other objects, so that it can only be accessed within the same object. If you want to allow external access to an attribute, you need to implement a getter or setter method for it.

But that’s not the only thing you can do to create a clean API. If you use no access modifiers for a class, it becomes package-private. This class can’t be accessed outside of its own package and is not part of the API. External clients of your software component are not aware of this class. They can only use it via a public class that uses the package-private class in a composition.

Let’s take a look at an example.

API design in the CoffeeMachine example

I use composition in the CoffeeMachine project that you might already know from the previous posts of this series. You can clone it at https://github.com/thjanssen/Stackify-OopAbstraction.

The CoffeeMachine class models a modern coffee machine with an integrated grinder and a brewing unit. In the real world, these two elements are parts of the coffee machine and can’t be separated. You also don’t interact with them directly. You always use them via the interface of the coffee machine. This interface only you gives you access to the operations that are required to brew a coffee and hides every other detail.

That design works pretty well in the real world. Let’s try the same approach for our example application.

Grinder and BrewingUnit as internal classes

The Grinder and BrewingUnit classes are package-private and can’t be accessed from the outside. You will not even see them, as long as you’re not adding a class to the package of my application.

class Grinder { 
    public GroundCoffee grind(CoffeeBean coffeeBean, double quantityCoffee) { 
        return new GroundCoffee(); 
    } 
}

class BrewingUnit { 
    public Coffee brew(CoffeeSelection selection, GroundCoffee groundCoffee, double quantity) { 
        return new Coffee(selection, quantity); 
    } 
}

The CoffeeMachine class defines the public API

The public methods of the CoffeeMachine class define the main part of the API of my small sample application. These are:

• a constructor that expects a Map of coffee beans,
• the addBeans method which enables you to refill coffee beans or to add different ones
• the brewCoffee method that you can call with a CoffeeSelection to brew a cup of filter coffee or espresso

import java.util.HashMap; 
import java.util.Map; 
 
public class CoffeeMachine { 
    private Map<CoffeeSelection, Configuration> configMap; 
    private Map<CoffeeSelection, CoffeeBean> beans; 
    private Grinder grinder; 
    private BrewingUnit brewingUnit; 
 
    public CoffeeMachine(Map<CoffeeSelection, CoffeeBean> beans) { 
        this.beans = beans; 
        this.grinder = new Grinder(); 
        this.brewingUnit = new BrewingUnit(); 
 
        this.configMap = new HashMap<CoffeeSelection, Configuration>(); 
        this.configMap.put(CoffeeSelection.ESPRESSO, new Configuration(8, 28)); 
        this.configMap.put(CoffeeSelection.FILTER_COFFEE, new Configuration(30, 480)); 
    } 
 
    public Coffee brewCoffee(CoffeeSelection selection) throws CoffeeException { 
        switch (selection) { 
            case FILTER_COFFEE: 
                return brewFilterCoffee(); 
            case ESPRESSO: 
                return brewEspresso(); 
            default: 
                throw new CoffeeException("CoffeeSelection [" + selection + "] not supported!"); 
        } 
    }   
 
    private Coffee brewEspresso() { 
        Configuration config = configMap.get(CoffeeSelection.ESPRESSO); 
 
        // grind the coffee beans 
        GroundCoffee groundCoffee = this.grinder.grind(
            this.beans.get(CoffeeSelection.ESPRESSO), config.getQuantityCoffee()); 
 
        // brew an espresso 
        return this.brewingUnit.brew(
            CoffeeSelection.ESPRESSO, groundCoffee, config.getQuantityWater()); 
    } 
 
    private Coffee brewFilterCoffee() { 
        Configuration config = configMap.get(CoffeeSelection.FILTER_COFFEE); 
 
        // grind the coffee beans 
        GroundCoffee groundCoffee = this.grinder.grind(
            this.beans.get(CoffeeSelection.FILTER_COFFEE), config.getQuantityCoffee()); 
 
        // brew a filter coffee 
        return this.brewingUnit.brew(
            CoffeeSelection.FILTER_COFFEE, groundCoffee, config.getQuantityWater()); 
    } 
 
    public void addBeans(CoffeeSelection sel, CoffeeBean newBeans)
        throws CoffeeException { 
        CoffeeBean existingBeans = this.beans.get(sel); 
        if (existingBeans != null) { 
            if (existingBeans.getName().equals(newBeans.getName())) { 
                existingBeans.setQuantity(existingBeans.getQuantity() + newBeans.getQuantity()); 
            } else { 
                throw new CoffeeException(
                    "Only one kind of beans supported for each CoffeeSelection."); 
            } 
        } else { 
            this.beans.put(sel, newBeans); 
        } 
    }
}

The Grinder and BrewingUnit objects can’t be accessed outside of their package. So, I need to instantiate them within the CoffeeMachine class. As you can see in the code snippet, I do that in the constructor method. This approach also allows you to control the usage of these objects used within the composition.

I use both of them in the brewFilterCoffee and the brewEspresso methods. That allows me to reuse the code provided by the two classes without exposing them to any client. And it also prevents any misuse of these objects. The CoffeeMachine class can validate the CoffeeBeans that are provided to the grind method of the Grinder and the kind of coffee and quantity of water that gets used to brew a selected coffee.

As you can see, composition allows you to design an API that’s easy and safe to use by encapsulating the details of your composition.

Hide internal code changes

Using composition and encapsulation not only enables you to create better APIs, but you can also use it to make your code easier to maintain and modify. As long as a class gets only used by your own code, you can easily change it and adapt any client code if necessary.

Change internal classes without side effects

The Grinder class in the CoffeeMachine example is not visible to any external clients. So, I could decide to change the signature of the grind method, or I could add additional methods to the Grinder class without worrying about any external side effects.

Let’s do that and add a CoffeeSelection as another parameter to the grind method. The Grinder can then select different coarseness settings for a filter coffee and an espresso.

class Grinder { 
    public GroundCoffee grind(CoffeeBean coffeeBean, double quantityCoffee, CoffeeSelection selection) { 
        return new GroundCoffee(selection); 
    } 
}

In this example, it’s just a simple change to the Grinder class. But that’s only the case because that class is package-private and used in a composition that doesn’t leak any information. I don’t need to worry about backward compatibility or how my changes might affect any code that uses the Grinder class.

Handle the changes internally

I just need to adjust the CoffeeMachine class because I know it’s the only class that uses the Grinder. The required change is simple. I just change the calls of the grind method in the brewFilterCoffee and brewEspresso methods and provide a CoffeeSelection as the third parameter.

import java.util.HashMap; 
import java.util.Map; 
 
public class CoffeeMachine { 
    private Map<CoffeeSelection, Configuration> configMap; 
    private Map<CoffeeSelection, CoffeeBean> beans; 
    private Grinder grinder; 
    private BrewingUnit brewingUnit; 
 
    public CoffeeMachine(Map<CoffeeSelection, CoffeeBean> beans) { 
        this.beans = beans; 
        this.grinder = new Grinder(); 
        this.brewingUnit = new BrewingUnit(); 
 
        this.configMap = new HashMap<CoffeeSelection, Configuration>(); 
        this.configMap.put(CoffeeSelection.ESPRESSO, new Configuration(8, 28)); 
        this.configMap.put(CoffeeSelection.FILTER_COFFEE, 
            new Configuration(30, 480)); 
    } 
 
    public Coffee brewCoffee(CoffeeSelection selection) throws CoffeeException { 
        switch (selection) { 
            case FILTER_COFFEE: 
                return brewFilterCoffee(); 
            case ESPRESSO: 
                return brewEspresso(); 
            default: 
                throw new CoffeeException(
                "CoffeeSelection [" + selection + "] not supported!");
        } 
    } 
 
    private Coffee brewEspresso() { 
        Configuration config = configMap.get(CoffeeSelection.ESPRESSO); 
 
        // grind the coffee beans 
        GroundCoffee groundCoffee = this.grinder.grind(
            this.beans.get(CoffeeSelection.ESPRESSO), config.getQuantityCoffee(), CoffeeSelection.ESPRESSO); 
 
        // brew an espresso 
        return this.brewingUnit.brew(
            CoffeeSelection.ESPRESSO, groundCoffee, config.getQuantityWater()); 
    } 
 
    private Coffee brewFilterCoffee() { 
        Configuration config = configMap.get(CoffeeSelection.FILTER_COFFEE); 
 
        // grind the coffee beans 
        GroundCoffee groundCoffee = this.grinder.grind(
            this.beans.get(CoffeeSelection.FILTER_COFFEE),
            config.getQuantityCoffee(), CoffeeSelection.FILTER_COFFEE); 
 
        // brew a filter coffee 
        return this.brewingUnit.brew(
            CoffeeSelection.FILTER_COFFEE, groundCoffee, config.getQuantityWater()); 
    } 
 
    public void addBeans(CoffeeSelection sel, CoffeeBean newBeans)
        throws CoffeeException { 
        CoffeeBean existingBeans = this.beans.get(sel); 
        if (existingBeans != null) { 
            if (existingBeans.getName().equals(newBeans.getName())) { 
                existingBeans.setQuantity(existingBeans.getQuantity() + newBeans.getQuantity()); 
            } else { 
                throw new CoffeeException(
                    "Only one kind of beans supported for each CoffeeSelection."); 
            } 
        } else { 
            this.beans.put(sel, newBeans); 
        } 
    } 
}

As you can see, the change of the Grinder class doesn’t have any effect on the API. The composition and encapsulation provided by the CoffeeMachine class hide the changes. That makes them a lot easier to implement and improves the maintainability of the example project.

Summary

Composition is one of the key concepts of object-oriented programming languages like Java. It enables you to reuse code by modeling a has-a association between objects.

If you combine the concept of composition with the encapsulation concept, you can exclude the reused classes from your API. That enables you to implement software components that are easy to use and maintain.

With APM, server health metrics, and error log integration, improve your application performance with Stackify Retrace.  Try your free two week trial today

Author

Thorben

More articles by Thorben