Massimo Caliman
by Massimo Caliman
2 min read


  • Programming

You can find an example project on github at this address:

Some information on this pattern.

It is a behavioural design pattern (Behavioural patterns provide solutions to the most common types of interaction between objects)

It is used when an object must change its behaviour (at run-time) according to its internal state.

Avoid using if-else conditional block sequences ( AntiIF Campaign )

It provides a systematic and lose-coupled way to achieve this.

It consists of a State interface, a concrete class Context and one or more concrete classes implementing the State interface.

  • Context defines the client class interface and holds a reference to a ConcreteState.
  • State defines the interface implemented by ConcreteState.
  • ConcreteState classes encapsulate the logic of the behaviour associated with a given state.

Let’s see how it works with a brief example on github and the code of which I give in this post.

First we define the interface, it only needs to have an execute method.

public interface State {
   public void execute();

We want to implement a content management system with three states: draft, published and deleted.

Our test application may have three buttons to perform the respective actions.

The solution that adopts the State design pattern is one that implements as many concrete State classes as there are states.

public class DarftState implements State {
   public void execute() {        

Let’s create the Context class, which must have a reference to an object of type State. An object of type Context will need to have a mechanism for changing the state attribute and a method that propagates the action performed by the execute method to it.

public class Context {
   private State state;
   public void setState(State state) {
      this.state = state;
   public State getState() {
      return this.state;
   public void execute() {

The Context class in our case could also be called Post and contain multiple attributes and methods, e.g. containing the reference to the record or content entity whose lifecycle we want to manage. For the other states I refer you to the code on github whose references are at the top of this post. Finally, an example of use that changes the state of our object three times.

Context context = new Context();
State darftState = new DarftState();
State publishState = new PublishState();
State deletedState = new DeletedState();




If a new behaviour or state is to be added, we must design a new class that implements State.

What disadvantages can there be in adopting this pattern? It obviously increases the number of classes, but this is preferable in most cases to a single class with long sequences of conditional if (or switch) statements. A not inconsiderable advantage: in a team of developers writing individual ConcreteStates can be assigned and managed easily compared to the alternative of
getting your hands on two or more of the same class.