Interfaces

Interfaces can provide a layer of abstraction to your code. They separate the specific implementation of a method from the declaration for that method. This way you can have different implementations of a method based on your specific application.

Defining an interface is similar to defining a new class. For example, a company can have two types of purchase orders, ones that come from customers, and others that come from their employees. Both are a type of purchase order. Suppose you needed a method to provide a discount. The amount of the discount can depend on the type of purchase order.

You can model the general concept of a purchase order as an interface and have specific implementations for customers and employees. In the following example the focus is only on the discount aspect of a purchase order.

Here’s the definition of the PurchaseOrder interface.

1// An interface that defines what a purchase order looks like in general
2public interface PurchaseOrder {
3    // All other functionality excluded
4    Double discount();
5}

This class implements the PurchaseOrder interface for customer purchase orders.

1// One implementation of the interface for customers
2public class CustomerPurchaseOrder implements PurchaseOrder {
3    public Double discount() {
4        return .05;  // Flat 5% discount
5    }
6}

This class implements the PurchaseOrder interface for employee purchase orders.

1// Another implementation of the interface for employees
2public class EmployeePurchaseOrder implements PurchaseOrder {
3      public Double discount() {
4        return .10;  // It’s worth it being an employee! 10% discount
5      } 
6}

When you define a new interface, you’re defining a new data type. You can use an interface name in any place you can use another data type name. Any object assigned to a variable of type interface must be an instance of a class that implements the interface, or a sub-interface data type.

See also Classes and Casting.