WCF Part 2 : Defining contract
As seen in the previous article, we need an address, binding and contract to complete the WCF ABC. We'll start at the contract.
A contract is defined explicitly, via a class. You add a [ServiceContract] attribute to the class. All methods you want to expose in your service, you mark as [OperationContract], as methods are called operations in services.
[ServiceContract]
public class Hello
{
[OperationContract]
string HelloWorld()
{
return "Hello world";
}
string HelloComputer()
{
return "Hello computer";
}
}
The operations are defined explicitly, a Service Orientation tenet. HelloComputer won't be visible by consumers of our service; it isn't marked with an attribute. The HelloWorld operation however is, even though it's private inside the .NET World.
Interfaces
We prefer however, to have an interface for our contract and have our actual service implement the interface. That's because
- Interfaces can extend/inherit other interfaces
- A single class can implement multiple interfaces
- You can modify the implementation of a service without breaking the contract
- You can version your service via old and new interfaces
It's always best to have an interface and implement it. The attributes also must be specified in the interface.
[ServiceContract]
public interface IHello
{
[OperationContract]
string HelloWorld();
string HelloComputer();
}
public class Hello : IHello
{
string IHello.HelloWorld()
{
return "Hello world";
}
string IHello.HelloComputer()
{
return "Hello computer";
}
}
In the next article I'll show you how you can host this service. In the future we'll also consume the service and I will try to tell more about contracts and versioning.