Programming iOS 4

Interface and Implementation

As you already know from Chapter 2, all your code is going to go into some class or other. So the first thing we must do is specify what is meant by putting code “into a class” in Objective-C. How does Objective-C say, linguistically and structurally, “This is the code for such-and-such a class”?

To write the code for class, you must actually provide two chunks or sections of code, called the interface and the implementation. Here’s the complete minimum code required to define a class called MyClass. This class is so minimal that it doesn’t even have any methods of its own:

@interface MyClass
@end
@implementation MyClass
@end

The @interface and @implementation compiler directives show the compiler where the interface and implementation sections begin for the class that’s being defined, MyClass; the corresponding @end lines show where each of those sections end.

In real life, the implementation section is where any methods for MyClass would be defined. So here’s a class that’s actually defined to do something:

@interface MyClass
@end
@implementation MyClass
- (NSString*) sayGoodnightGracie {
    return @"Good night, Gracie!";
}
@end

Observe how a method is defined. The first line is just like the method declaration, stating the type of method (class or instance), the type of value returned, and the name of the method along with the types of any parameters and local names for those parameters (see Chapter 3). Then come curly braces containing the code to be executed when the method is called, just as with a C function (see Chapter 1).

However, this class is still pretty much useless, because it can’t be instantiated. In Cocoa, knowledge of how to be instantiated, plus how to do a number of other things that any class should know how to do, resides in the base class, which is the NSObject class. Therefore, all Cocoa classes must be based ultimately on the NSObject class, by declaring as the superclass for your class either NSObject or some other class that inherits from NSObject (as just about any other Cocoa class does). The syntax for this declaration is a colon followed by the superclass name in the @interface line, like this:

@interface MyClass : NSObject
@end
@implementation MyClass
- (NSString*) sayGoodnightGracie {
    return @"Good night, Gracie!";
}
@end

Note

NSObject is not the only Cocoa base class. It used to be, but there is now another, NSProxy. NSProxy is used only in very special circumstances and is not discussed in this book. If you have no reason for your class to inherit from any other class, make it inherit from NSObject.

In its fullest form, the interface section might contain some more material. In particular, there are two main types of stuff that the interface section might contain:

Instance variables

If our class is to have any instance variables (other than those inherited from its superclass), they must be declared in the interface section.

Method declarations

If we want to declare our methods, those method declarations go into the interface section. Method declarations are not required, but without a method declaration, a method cannot be “seen” by other methods defined before it in the same class.

So here is MyClass defined in what we might term canonical form:

@interface MyClass : NSObject {
    // instance variable declarations go here
}
- (NSString*) sayGoodnightGracie;
@end
@implementation MyClass
- (NSString*) sayGoodnightGracie {
    return @"Good night, Gracie!";
}
@end

There are no instance variable declarations in our class, so I’ve used a comment to show where they go; notice the curly braces surrounding the place. I’ll go into detail about instance variables in the next chapter. The method declaration matches the name and signature for the method definition and ends with a semicolon (required).