This chapter is from the book
This chapter is from the book
This chapter is from the book
When you define a new class, you have to do a few things. First, you have to tell the Objective-C compiler where the class came from. That is, you have to name its parent class. Second, you have to specify what type of data is to be stored in the objects of this class. That is, you have to describe the data that members of the class will contain. These members are called the instance variables. Finally, you need to define the type of operations, or methods, that can be used when working with objects from this class. This is all done in a special section of the program called the @interface section. The general format of this section looks like this:
@interface NewClassName: ParentClassName
{
memberDeclarations;
}
methodDeclarations;
@end
By convention, class names begin with an uppercase letter, even though it's not required. This enables someone reading your program to distinguish class names from other types of variables by simply looking at the first character of the name. Let's take a short diversion to talk a little bit about forming names in Objective-C.
Choosing Names
In Chapter 2, "Programming in Objective-C," you used several variables to store integer values. For example, you used the variable sum in Program 2.4 to store the result of the addition of the two integers, 50 and 25.
The Objective-C language allows data types other than just integers to be stored in variables as well, provided the proper declaration for the variable is made before it is used in the program. Variables can be used to store floating-point numbers, characters, and even objects (or more precisely, references to objects).
The rules for forming names are quite simple: They must begin with a letter or underscore (_), and they can be followed by any combination of letters (upper- or lowercase), underscores, or the digits 0–9. The following is a list of valid names:
sum
pieceFlag
i
myRectangle
numberOfMoves
_sysFlag
ChessPiece
On the other hand, the following names are not valid for the stated reasons:
sum$value—$ is not a valid character.
piece flag—Embedded spaces are not permitted.
3Spencer—Names can't start with a number.
int—This is a reserved word.
int cannot be used as a variable name because its use has a special meaning to the Objective-C compiler. This use is known as a reserved name or reserved word. In general, any name that has special significance to the Objective-C compiler cannot be used as a variable name. Appendix B, "Objective-C Language Summary," provides a complete list of such reserved names.
You should always remember that upper- and lowercase letters are distinct in Objective-C. Therefore, the variable names sum, Sum, and SUM each refer to a different variable. As noted, when naming a class, start it with a capital letter. Instance variables, objects, and method names, on the other hand, typically begin with lowercase letters. To aid readability, capital letters are used inside names to indicate the start of a new word, as in the following examples:
AddressBook This could be a class name.
currentEntry This could be an object.
current_entry Underscores are also used by some programmers as word separators.
addNewEntry This could be a method name.
When deciding on a name, keep one recommendation in mind: Don't be lazy. Pick names that reflect the intended use of the variable or object. The reasons are obvious. Just as with the comment statement, meaningful names can dramatically increase the readability of a program and will pay off in the debug and documentation phases. In fact, the documentation task will probably be greatly reduced because the program will be more self-explanatory.
Here again is the @interface section from Program 3.2:
//------- @interface section -------
@interface Fraction: Object
{
int numerator;
int denominator;
}
-(void) print;
-(void) setNumerator: (int) n;
-(void) setDenominator: (int) d;
@end
The name of the new class (NewClassName) is Fraction, and its parent class is Object. (We'll talk in greater detail about parent classes in Chapter 8, "Inheritance.") The Object class is defined in the file objc/Object.h, which is why that file (along with stdio.h) is imported at the beginning of Program 3.2:
#import <objc/Object.h>
Instance Variables
The memberDeclarations section specifies what types of data are stored in a Fraction, as well as the names of those data types. As you can see, this section is enclosed inside its own set of curly braces. For your Fraction class, the declarations
int numerator; int denominator;
say that a Fraction object has two integer members called numerator and denominator.
The members declared in this section are known as the instance variables. As you'll see, each time you create a new object, a new and unique set of instance variables also is created. Therefore, if you have two Fractions, one called fracA and another called fracB, each will have its own set of instance variables. That is, fracA and fracB each will have its own separate numerator and denominator. The Objective-C system automatically keeps track of this for you, which is one of the nicer things about working with objects.
Class and Instance Methods
You have to define methods to work with your Fractions. You'll need to be able to set the value of a fraction to a particular value. Because you won't have direct access to the internal representation of a fraction (in other words, direct access to its instance variables), you must write methods to set the numerator and denominator. You'll also write a method called print that will display the value of a fraction. Here's what the declaration for the print method looks like in the interface file:
-(void) print;
The leading minus sign (-) tells the Objective-C compiler that the method is an instance method. The only other option is a plus sign (+), which indicates a class method. A class method is one that performs some operation on the class itself, such as creating a new instance of the class. This is similar to manufacturing a new car, in that the car is the class and you want to create a new one—which would be a class method.
An instance method performs some operation on a particular instance of a class, such as setting its value, retrieving its value, displaying its value, and so on. Referring to the car example, after you have manufactured the car, you might need to fill it with gas. The operation of filling it with gas is performed on a particular car, so it would be analogous to an instance method.
Return Values
When you declare a new method, you have to tell the Objective-C compiler whether the method returns a value, and if it does, what type of value it returns. This is done by enclosing the return type in parentheses after the leading minus or plus sign. So, the declaration
-(int) getNumerator;
specifies that the instance method called getNumerator returns an integer value. Similarly, the line
-(double) getDoubleValue;
declares a method that returns a double precision value. (You'll learn more about this data type in Chapter 4, "Data Types and Expressions.")
A value is returned from a method using the Objective-C return statement, similar to the way in which we returned a value from main in previous program examples.
If the method returns no value, you indicate that using the type void, as in the following:
-(void) print;
This declares an instance method called print that returns no value. In such a case, you do not need to execute a return statement at the end of your method. Alternatively, you can execute a return without any specified value, as in the following:
return;
You don't need to specify a return type for your methods, although it's better programming practice if you do. If none is specified, id is the default. You'll learn more about the id data type in a later chapter. Basically, the id type can be used to refer to any type of object.
Method Arguments
Two other methods are declared in the @interface section from Program 3.2:
-(void) setNumerator: (int) n; -(void) setDenominator: (int) d;
These are both instance methods that return no value. Each method takes an integer argument, which is indicated by the (int) in front of the argument name. In the case of setNumerator, the name of the argument is n. This name is arbitrary, and it is the name the method uses to refer to the argument. Therefore, the declaration of setNumerator specifies that one integer argument called n will be passed to the method and that no value is returned. This is similar for setDenominator, except the name of its argument is d.
Notice the syntax of the declaration for these methods. Each method name ends with a colon, which tells the Objective-C compiler that the method expects to see an argument. Next, the type of the argument is specified, enclosed in a set of parentheses, in much the same way the return type is specified for the method itself. Finally, the symbolic name to be used to identify that argument in the method is specified. The entire declaration is terminated with a semicolon. This syntax is depicted in Figure 3.2.
){kind=link}
Figure
3.2 Declaring a method.
When a method takes an argument, you also append a colon to the method name when referring to the method. setNumerator: and setDenominator: would therefore be the correct way to identify these two methods—each of which takes a single argument. Also, the identification of the print method without a trailing colon indicates that this method does not take any arguments. In Chapter 7, "More on Classes," you'll see how methods that take more than one argument are identified.