About Jose Luis

Jose Luis is a computer scientist and professional senior software engineer with experience in a wide variety of technologies, particularly those oriented towards the JVM platform. He is mainly interested in back-end development and Agile practices that make software development both more fun and more valuable to customers.

Static factory methods vs traditional constructors

I’ve previously talked a little bit about the Builder Pattern, a useful pattern to instantiate classes with several (possibly optional) attributes that results in easier to read, write and maintain client code, among other benefits. Today, I’m going to continue exploring object creation techniques but this time for a more general case.

Take the following example, which is by no means a useful class other than to make my point. We have a RandomIntGenerator class that, as the name suggests, generates random int numbers. Something like:
 
 
 

public class RandomIntGenerator {
    private final int min;
    private final int max;

    public int next() {...}
}

Our generator takes a minimum and maximum and then generates random numbers between those 2 values. Notice that the two attributes are declared final so we have to initialize them either on their declaration or in the class constructor. Let’s go with the constructor:

    public RandomIntGenerator(int min, int max) {
        this.min = min;
        this.max = max;
    }

Now, we also want to give our clients the possibility to specify just a minimum value and then generate random values between that minimum and the max possible value for ints. So we add a second constructor:

    public RandomIntGenerator(int min) {
        this.min = min;
        this.max = Integer.MAX_VALUE;
    }

So far so good, right? But in the same way that we provided a constructor to just specify the minimum value, we want to do the same for just the maximum. We’ll just add a third constructor like:

    public RandomIntGenerator(int max) {
        this.min = Integer.MIN_VALUE;
        this.max = max;
    }

If you try that, you’ll get a compilation error that goes: Duplicate method RandomIntGenerator(int) in type RandomIntGenerator. What’s wrong? The problem is that constructors, by definition, have no names. As such, a class can only have one constructor with a given signature in the same way that you can’t have two methods with the same signature (same return type, name and parameters type). That is why when we tried to add the RandomIntGenerator(int max) constructor we got that compilation error, because we already had the RandomIntGenerator(int min) one.

Is there something we can do in cases like this one? Not with constructors but fortunately there’s something else we can use: static factory methods, which are simply public static methods that return an instance of the class. You’ve probably used this technique without even realizing it. Have you ever used Boolean.valueOf? It looks something like:

    public static Boolean valueOf(boolean b) {
        return (b ? TRUE : FALSE);
    }

Applying static factories to our RandomIntGenerator example, we could get:

public class RandomIntGenerator {
    private final int min;
    private final int max;

    private RandomIntGenerator(int min, int max) {
        this.min = min;
        this.max = max;
    }

    public static RandomIntGenerator between(int max, int min) {
        return new RandomIntGenerator(min, max);
    }

    public static RandomIntGenerator biggerThan(int min) {
        return new RandomIntGenerator(min, Integer.MAX_VALUE);
    }

    public static RandomIntGenerator smallerThan(int max) {
        return new RandomIntGenerator(Integer.MIN_VALUE, max);
    }

    public int next() {...}
}

Note how the constructor was made private to ensure that the class is only instantiated through its public static factory methods. Also note how your intent is clearly expressed when you have a client with RandomIntGenerator.between(10,20) instead of new RandomIntGenerator(10,20). It’s worth mentioning that this technique is not the same as the Factory method Design Pattern from the Gang of Four. Any class can provide static factory methods instead of, or in addition to, constructors. So what are the advantages and disadvantages of this technique? We already mentioned the first advantage of static factory methods: unlike constructors they have names. This has two direct consequences,

  1. We can provide a meaningful name for our constructors.
  2. We can provide several constructors with the same number and type of parameters, something that as we saw earlier we can’t do with class constructors.

Another advantage of static factories is that, unlike constructors, they are not required to return a new object every time they are invoked. This is extremely useful when working with immutable classes to provide constant objects for common used values and avoid creating unnecessary duplicate objects. The Boolean.valueOf code that I showed previously illustrates this point perfectly. Notice that this static method returns either TRUE or FALSE, both immutable Boolean objects.

A third advantage of static factory methods is that they can return an object of any subtype of their return type. This gives you the possibility to change the return type freely without affecting clients. Moreover, you can hide implementation classes and have an interface-based API, which is usually a really good idea. But I think this can be better seen by an example.

Remember the RandomIntGenerator at the beginning of this post? Let’s make that a little bit more complicated. Imagine that we now want to provide random generators not just for integers but for other data-types like String, Double or Long. They are all going to have a next() method that returns a random object of a particular type, so we could start with an interface like:

public interface RandomGenerator<T> {
    T next();
}

Our first implementation of the RandomIntGenerator now becomes:

class RandomIntGenerator implements RandomGenerator<Integer> {
    private final int min;
    private final int max;

    RandomIntGenerator(int min, int max) {
        this.min = min;
        this.max = max;
    }

    public Integer next() {...}
}

We could also have a String generator:

class RandomStringGenerator implements RandomGenerator<String> {
    private final String prefix;

    RandomStringGenerator(String prefix) {
        this.prefix = prefix;
    }

    public String next() {...}
}

Notice how all the classes are declared package-private (default scope) and so are their constructors. This means that no client outside of their package can create instances of these generators. So what do we do? Tip: It starts with “static” and ends with “methods”.
Consider the following class:

public final class RandomGenerators {
    // Suppresses default constructor, ensuring non-instantiability.
    private RandomGenerators() {}

    public static final RandomGenerator<Integer> getIntGenerator() {
        return new RandomIntGenerator(Integer.MIN_VALUE, Integer.MAX_VALUE);
    }

    public static final RandomGenerator<String> getStringGenerator() {
        return new RandomStringGenerator('');
    }
}

RandomGenerators is just a noninstantiable utility class with nothing else than static factory methods. Being on the same package as the different generators this class can effectively access and instantiate those classes. But here comes the interesting part. Note that the methods only return the RandomGenerator interface, and that’s all the clients need really. If they get a RandomGenerator<Integer> they know that they can call next() and get a random integer.
Imagine that next month we code a super efficient new integer generator. Provided that this new class implements RandomGenerator<Integer> we can change the return type of the static factory method and all clients are now magically using the new implementation without them even noticing the change.

Classes like RandomGenerators are quite common both on the JDK and on third party libraries. You can see examples in Collections (in java.util), Lists, Sets or Maps in Guava. The naming convention is usually the same: if you have an interface named Type you put your static factory methods in a noninstantiable class named Types.

A final advantage of static factories is that they make instantiating parameterized classes a lot less verbose. Have you ever had to write code like this?

Map<String, List<String>> map = new HashMap<String, List<String>>();

You are repeating the same parameters twice on the same line of code. Wouldn’t it be nice if the right side of the assign could be inferred from the left side? Well, with static factories it can. The following code is taken from Guava’s Maps class:

  public static <K, V> HashMap<K, V> newHashMap() {
    return new HashMap<K, V>();
  }

So now our client code becomes:

Map<String, List<String>> map = Maps.newHashMap();

Pretty nice, isn’t it? This capability is known as Type inference. It’s worth mentioning that Java 7 introduced type inference through the use of the diamond operator. So if you’re using Java 7 you can write the previous example as:

Map<String, List<String>> map = new HashMap<>();

The main disadvantage of static factories is that classes without public or protected constructors cannot be extended. But this might be actually a good thing in some cases because it encourages developers to favor composition over inheritance.

To summarize, static factory methods provide a lot of benefits and just one drawback that might actually not be a problem when you think about it. Therefore, resist the urge to automatically provide public constructors and evaluate if static factories are a better fit for your class.
 

Reference: Static factory methods vs traditional constructors from our JCG partner Jose Luis at the Development the way it should be blog.

Related Whitepaper:

Bulletproof Java Code: A Practical Strategy for Developing Functional, Reliable, and Secure Java Code

Use Java? If you do, you know that Java software can be used to drive application logic of Web services or Web applications. Perhaps you use it for desktop applications? Or, embedded devices? Whatever your use of Java code, functional errors are the enemy!

To combat this enemy, your team might already perform functional testing. Even so, you're taking significant risks if you have not yet implemented a comprehensive team-wide quality management strategy. Such a strategy alleviates reliability, security, and performance problems to ensure that your code is free of functionality errors.Read this article to learn about this simple four-step strategy that is proven to make Java code more reliable, more secure, and easier to maintain.

Get it Now!  

Leave a Reply


× 6 = forty eight



Java Code Geeks and all content copyright © 2010-2014, Exelixis Media Ltd | Terms of Use
All trademarks and registered trademarks appearing on Java Code Geeks are the property of their respective owners.
Java is a trademark or registered trademark of Oracle Corporation in the United States and other countries.
Java Code Geeks is not connected to Oracle Corporation and is not sponsored by Oracle Corporation.

Sign up for our Newsletter

15,153 insiders are already enjoying weekly updates and complimentary whitepapers! Join them now to gain exclusive access to the latest news in the Java world, as well as insights about Android, Scala, Groovy and other related technologies.

As an extra bonus, by joining you will get our brand new e-books, published by Java Code Geeks and their JCG partners for your reading pleasure! Enter your info and stay on top of things,

  • Fresh trends
  • Cases and examples
  • Research and insights
  • Two complimentary e-books
Get tutored by the Geeks! JCG Academy is a fact... Join Now
Hello. Add your message here.