1. Overview

In this tutorial, we’ll describe the Observer pattern and take a look at a few Java implementation alternatives.

2. What Is the Observer Pattern?

Observer is a behavioral design pattern. It specifies communication between objects: observable and observers. An observable is an object which notifies observers about the changes in its state.

For example, a news agency can notify channels when it receives news. Receiving news is what changes the state of the news agency, and it causes the channels to be notified.

Let’s see how we can implement it ourselves.

First, we’ll define the NewsAgency class:

public class NewsAgency {
    private String news;
    private List<Channel> channels = new ArrayList<>();

    public void addObserver(Channel channel) {
        this.channels.add(channel);
    }

    public void removeObserver(Channel channel) {
        this.channels.remove(channel);
    }

    public void setNews(String news) {
        this.news = news;
        for (Channel channel : this.channels) {
            channel.update(this.news);
        }
    }
}

NewsAgency is an observable, and when news gets updated, the state of NewsAgency changes. When the change happens, NewsAgency notifies the observers about it by calling their update() method.

To be able to do that, the observable object needs to keep references to the observers. In our case, it’s the channels variable.

Now let’s see what the observer, the Channel class, can look like. It should have the update() method, which is invoked when the state of NewsAgency changes:

public class NewsChannel implements Channel {
    private String news;

    @Override
    public void update(Object news) {
        this.setNews((String) news);
    } 

    // standard getter and setter
}

The Channel interface has only one method:

public interface Channel {
    public void update(Object o);
}

Now if we add an instance of NewsChannel to the list of observers, and change the state of NewsAgency, the instance of NewsChannel will be updated:

NewsAgency observable = new NewsAgency();
NewsChannel observer = new NewsChannel();

observable.addObserver(observer);
observable.setNews("news");
assertEquals(observer.getNews(), "news");

There’s a predefined Observer interface in Java core libraries, which makes implementing the observer pattern even more simple. Let’s look at it.

3. Implementation With Observer

The java.util.Observer interface defines the update() method, so there’s no need to define it ourselves, as we did in the previous section.

Let’s see how we can use it in our implementation:

public class ONewsChannel implements Observer {

    private String news;

    @Override
    public void update(Observable o, Object news) {
        this.setNews((String) news);
    }

    // standard getter and setter
}

Here, the second argument comes from Observable, as we’ll see below.

To define the observable, we need to extend Java’s Observable class:

public class ONewsAgency extends Observable {
    private String news;

    public void setNews(String news) {
        this.news = news;
        setChanged();
        notifyObservers(news);
    }
}

Note that we don’t need to call the observer’s update() method directly. We just call setChanged() and notifyObservers(), and the Observable class will do the rest for us.

It also contains a list of observers and exposes methods to maintain that list, addObserver() and deleteObserver().

To test the result, we just need to add the observer to this list and set the news:

ONewsAgency observable = new ONewsAgency();
ONewsChannel observer = new ONewsChannel();

observable.addObserver(observer);
observable.setNews("news");
assertEquals(observer.getNews(), "news");

The Observer interface isn’t perfect, and has been deprecated since Java 9. One of the cons is that Observable isn’t an interface, it’s a class, and that’s why subclasses can’t be used as observables.

Also, a developer could override some of Observable‘s synchronized methods and disrupt their thread-safety.

Now let’s look at the ProperyChangeListener interface, which is recommended over Observer.

4. Implementation With PropertyChangeListener

In this implementation, an observable must keep a reference to the PropertyChangeSupport instance. It helps to send the notifications to observers when a property of the class is changed.

Let’s define the observable:

public class PCLNewsAgency {
    private String news;

    private PropertyChangeSupport support;

    public PCLNewsAgency() {
        support = new PropertyChangeSupport(this);
    }

    public void addPropertyChangeListener(PropertyChangeListener pcl) {
        support.addPropertyChangeListener(pcl);
    }

    public void removePropertyChangeListener(PropertyChangeListener pcl) {
        support.removePropertyChangeListener(pcl);
    }

    public void setNews(String value) {
        support.firePropertyChange("news", this.news, value);
        this.news = value;
    }
}

Using this support, we can add and remove observers, and notify them when the state of the observable changes:

support.firePropertyChange("news", this.news, value);

Here, the first argument is the name of the observed property. The second and the third arguments are its old and new value, accordingly.

Observers should implement PropertyChangeListener:

public class PCLNewsChannel implements PropertyChangeListener {

    private String news;

    public void propertyChange(PropertyChangeEvent evt) {
        this.setNews((String) evt.getNewValue());
    }

    // standard getter and setter
}

Due to the PropertyChangeSupport class, which is doing the wiring for us, we can restore the new property value from the event.

Let’s test the implementation to make sure that it also works:

PCLNewsAgency observable = new PCLNewsAgency();
PCLNewsChannel observer = new PCLNewsChannel();

observable.addPropertyChangeListener(observer);
observable.setNews("news");

assertEquals(observer.getNews(), "news");

5. Lapsed Listener Problem

All the above implementations require explicit observers’ deregistration to avoid the Lapsed Listener Problem. Forgetting to do so might produce a memory leak.

The best approach to resolve this issue is using WeakReferences, so when an observer goes out of the scope, it will be automatically removed from the observers list.

6. Conclusion

In this article, we examined two ways to implement the Observer design pattern in Java. We learned that the PropertyChangeListener approach is preferred.

The source code for this article is available over on GitHub.

Course – LS (cat=Spring)

Get started with Spring and Spring Boot, through the Learn Spring course:

>> THE COURSE
res – REST with Spring (eBook) (everywhere)
Comments are closed on this article!