1. Introduction
In this quick tutorial, we’ll learn about marker interfaces in Java.
2. Marker Interfaces
A marker interface is an interface that doesn’t have any methods or constants inside it. It provides run-time type information about objects, so the compiler and JVM have additional information about the object.
A marker interface is also called a tagging interface.
Though marker interfaces are still in use, they very likely point to a code smell, and we should use them carefully. The main reason for this is that they blur the lines of what an interface represents, since markers don’t define any behavior. Newer development favors annotations to solve some of the same problems.
3. JDK Marker Interfaces
Java has many built-in marker interfaces, such as Serializable, Cloneable, and Remote.
Let’s take the example of the Cloneable interface. If we try to clone an object that doesn’t implement this interface, the JVM throws a CloneNotSupportedException. Thus, the Cloneable marker interface is an indicator to the JVM that we can call the Object.clone() method.
In the same way, when calling the ObjectOutputStream.writeObject() method, the JVM checks if the object implements the Serializable marker interface. When this isn’t the case, a NotSerializableException is thrown. Therefore, the object isn’t serialized to the output stream.
4. Custom Marker Interface
Let’s create our own marker interface.
For example, we can create a marker that indicates whether an object can be removed from the database:
public interface Deletable {
}
In order to delete an entity from the database, the object representing this entity has to implement our Deletable marker interface:
public class Entity implements Deletable {
// implementation details
}
Let’s say that we have a DAO object with a method for removing entities from the database. We can write our delete() method so that only objects implementing our marker interface can be deleted:
public class ShapeDao {
// other dao methods
public boolean delete(Object object) {
if (!(object instanceof Deletable)) {
return false;
}
// delete implementation details
return true;
}
}
As we can see, we’re giving an indication to the JVM about the runtime behavior of our objects. If the object implements our marker interface, it can be deleted from the database.
5. Marker Interfaces vs. Annotations
By introducing annotations, Java provided us with an alternative to achieve the same results as the marker interfaces. Moreover, like marker interfaces, we can apply annotations to any class, and we can use them as indicators to perform certain actions.
So what’s the key difference?
Unlike annotations, interfaces allow us to take advantage of polymorphism. As a result, we can add additional restrictions to the marker interface.
For instance, let’s add a restriction that only a Shape type can be removed from the database:
public interface Shape {
double getArea();
double getCircumference();
}
In this case, our marker interface, DeletableShape, will look like this:
public interface DeletableShape extends Shape {
}
Then our class will implement the marker interface:
public class Rectangle implements DeletableShape {
// implementation details
}
Therefore, all DeletableShape implementations are also Shape implementations. Obviously, we can’t do that using annotations.
However, every design decision has trade-offs, and we can use polymorphism as a counter-argument against marker interfaces. In our example, every class extending Rectangle will automatically implement DeletableShape.
6. Marker Interfaces vs. Typical Interfaces
In the previous example, we could get the same results by modifying our DAO’s delete() method to test whether our object is a Shape or not*,* instead of testing whether it’s a Deletable:
public class ShapeDao {
// other dao methods
public boolean delete(Object object) {
if (!(object instanceof Shape)) {
return false;
}
// delete implementation details
return true;
}
}
So why create a marker interface when we can achieve the same results using a typical interface?
Let’s imagine that, in addition to the Shape type, we want to remove the Person type from the database as well. In this case, there are two options to achieve that.
The first option is to add an additional check to our previous delete() method to verify whether or not the object to delete is an instance of Person:
public boolean delete(Object object) {
if (!(object instanceof Shape || object instanceof Person)) {
return false;
}
// delete implementation details
return true;
}
But what if we have more types that we want to remove from the database as well? Obviously, this won’t be a good option because we have to change our method for every new type.
The second option is to make the Person type implement the Shape interface, which acts as a marker interface. But is a Person object really a Shape? The answer is clearly no, and that makes the second option worse than the first one.
Consequently, although we can achieve the same results by using a typical interface as a marker, we’ll end up with a poor design.
7. Conclusion
In this article, we learned about marker interfaces and how to use them. Then we explored some built-in Java examples of this type of interface, and how the JDK uses them.
Next, we created our own marker interface and weighed it against using an annotation. Finally, we demonstrated why it’s good practice to use a marker interface in some scenarios instead of a traditional interface.
As always, the code can be found on GitHub.