Extension 9: Pointers

You’ve probably noticed that when we declare objects,

Fraction *myFraction;

We put an asterisk between the object type and the name of the object (it doesn’t matter if you place it after the type, like Fraction*, in the middle, like Fraction * myFraction, or as above). This is not a typo; we’re saying that we are creating a pointer to an object.

Why Pointers?

At the very basic level, objects are simply a collection of data (and methods). This data has to reside somewhere, and for the most part, this data is located in your computer’s RAM. But some objects can contain a lot of data, and start to take up a lot of RAM. This problem is further augmented if you have many objects. Of course, it’s always a good idea to minimize RAM usage, especially on a platform like iPhone where RAM can be scarce. Pointers, rather than storing an actual copy of an object, simply store the memory address of an object—it points to an object. Therefore, a pointer takes up a lot less RAM than an object itself would. This also makes code more efficient.

You’ll probably notice that in our methods, we also pass pointers around rather than actual objects (as specified by the asterisk following any class type). This means, first of all, that we are only passing small blocks of data around, not a large amount of data that corresponds to an object itself. This also means that you’re not passing in a copy of an object—you’re passing in a copy of a pointer. This will have significant effects, as described below. But first,

What are Pointers?

Pointers point to objects in memory

Pointers "point" to actual objects in memory

A pointer is simply a value that stores an in-memory location. They are generally written in hexadecimal, and look something like 0x0ab23ef8. You don’t have to worry about what they look like, or how they work. They are generally treated as objects, but they aren’t. There only exists one copy of the actual object in memory (unless you explicitly make a copy), but multiple pointers can point to that same object.

Multiple pointers can point to the same object

Multiple pointers pointing to the same object


The most important consideration here is memory management. Given the diagram above, consider what would happen if you wrote this line:

anotherFrac = myFraction;

This is a rather innocent line, and it does assign the value of myFraction to anotherFrac. But remember that myFraction contains a memory address, not an actual Fraction object. This means that now anotherFrac points to the same object as myFraction, and, more importantly, nothing points to the original anotherFrac:

If you just assign pointers, you could leak memory

Don't do this

Now, you have no way to get to the original anotherFrac, and so that object will reside forever in memory (or until the application is quit or the system is turned off). THis object is considered to be leaked; if you leak too many objects, you’ll run out of memory, and on the iPhone the system will force your app to quit. Regardless of where you are, leaking is not a good thing.

The other thing to consider is the case of multiple pointers to the same object, as you might have when you pass a pointer to an object as an argument to a method (as you usually do). Any changes that you make to one of the pointers will also be reflected in the other (because they point to the same object in memory, after all). This can have undesired consequences; you could also use this to your advantage. For example, if you wrote a class method that reduced a Fraction, you could have the method directly affect the original fraction you pass in; you could also make a copy of that fraction. On the other hand, you might mess up a data set from the user, forcing you prompt the user again for the same data…

Pointers are a central tenet of Objective-C (and all of object-oriented programming). All objects are represented as pointers—the original object is implicitly created. Used properly, they are a great tool.

Objective-C Lesson 3: Object-Oriented Programming

When object-oriented programming arrived in the programming world, it was considered a savior of software by some and yet another foolish experiment by others. Regardless of the perspective, we need to define some basic concepts and terminology before we begin

A Thing

An object is a thing. In the real world, you have “things,” all of which have certain properties. For example, you probably have a car—a “thing.” The car has properties, such as paint color, total mileage, make, model, and year. You can also do things with the car, or have something done to it—for example, you can drive twenty miles in the car, or wash the car.

In an object-oriented program, your specific car is an instance of a class, generally called Car. A class is the definition of the object, and in most cases can be thought of as a “factory” for its instances. A class can create many instances of the same type of object, although they have to have different names. An object is simply a variable, but it can be (and usually is) of a custom type.

Instance Variables and Methods

As mentioned before, your car, and any other car, has certain properties that are potentially unique to that specific instance. These properties are known as instance variables, or ivars for short. Instance variables are variables, of either the standard data types, or of any other object, but they are private to the object—that is, every single instance of Car has its own identical set of instance variables; although the variables may be of the same type and number, they can (and usually do) contain different values across different instances.

A method is the proper term for a function, or simply a piece of code that is called upon to perform some action. In Objective-C, there are two types of methods—class methods and instance methods. Class methods are performed directly on the class that an object belongs to—for example, a class methods might ask the car factory how many cars it produced the previous year. Instance methods are performed on specific instances of a class—for example, you might ask one specific car to drive twenty miles, or you might wash one specific car; you wouldn’t do those things to the entire factory. Instance methods affect the state or condition of the object. Just like two identical cars adopt different characteristics over their lifetime, different instances of Car can take on different traits as well.

Message Objects

In Objective-C, there is a specific syntax for calling methods:

[ClassOrInstance method];

The method call begins with an open bracket. This is followed by the name of the class or instance, depending on whether the method belongs to a class or instance. This is followed by a space, and then the name of the method. This is followed by a closing bracket, and the all-important semicolon.
In more formal terminology, the following message call might be described as

[receiver message];

The meaning is the same.

Sample Program

Now, with the theory cleared up, we can move on to our first program involving classes.

Program 3.1

// A simple implementation of a Fraction
#import <Foundation/Foundation.h>

int main (int argc, const char * argv[]) {
	NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init];

	int numerator = 2;
	int denominator = 5;
	NSLog(@"This is a fraction with a value of %d/%d", numerator, denominator);

	[pool drain];
	return 0;

The output is

This is a fraction with a value of 2/5

In Program 3.1, we defined a numerator and denominator variable, and set them to integer values. We then printed the result of this simple fraction.

This program is short and condensed, and seems to be pretty simple—and it is. But what if we had a hundred fractions that we wanted to store? We’d have to keep track of two hundred variables. And performing operations on those variables would not be any easier.

As a matter of fact, Program 3.1 is not a true object-oriented program. It is still written using C-style code. We have not created any objects in the code; our Fraction object still exists only psychologically.

Program 3.2 presents the exact same functionality, except that we actually define this object in code. We then refer to the object as Fraction throughout the program. Don’t worry if it looks daunting; an explanation follows.

Program 3.2

// A Fraction class
#import <Foundation/Foundation.h>

//———————— Interface ————————

@interface Fraction : NSObject
	int numerator;
	int denominator;

- (void)showResults;
- (void)setNumerator:(int)n;
- (void)setDenominator:(int)d;


//———————— Implementation ————————

@implementation Fraction

- (void)showResults {
	NSLog(@"This is a fraction with a value of %d/%d", numerator, denominator);

- (void)setNumerator:(int)n {
	numerator = n;

- (void)setDenominator:(int)d {
	denominator = d;


//———————— Main Program ————————

int main (int argc, const char * argv[]) {
    NSAutoreleasePool * pool = [[NSAutoreleasePool alloc] init];

    // Create an instance of Fraction
	Fraction *myFraction;

	myFraction = [[Fraction alloc] init];

	// Set myFraction to 2/5
	[myFraction setNumerator:2];
	[myFraction setDenominator:5];

	// Display the value of myFraction
	[myFraction showResults];

	// Clear up memory
	[myFraction release];

    [pool drain];
    return 0;

The output is

This is a fraction with a value of 2/5

Code, Demystified

Program 3.2 is divided into three logical sections. These are known as the interface, implementation, and program section, respectively.

The interface section (delineated by @interface in the code) provides basic program-wide definitions, the implementation section (delineated by @implementation) provides the implementation of the methods, and the program section contains the main code that does something.

The @interface Section

Every class in Objective-C has a single parent, or superclass, from which the class inherits. Next, you have to declare the instance variables that belong to the class. Finally, you declare the methods, or actions, that the class can perform. This is all done in the @interface section, which looks like this:

@interface ClassName : ParentClassName

Remember that the compiler ignores spaces, so you may put line breaks and spaces anywhere within the code above.
By convention, the name of a class begins with a capital letter, as in the Fraction class.

You can tell that our Fraction class is a subclass of NSObject, which is the root object that almost everything else inherits from, directly or indirectly. NSObject is declared in the Foundation files which you imported.

Fraction contains two instance variables; both are ints. They are called instance variables because every instance of Fraction creates a new set of these variables, unique to that specific instance. So, FractionA’s numerator is a different variable than FractionB’s numerator. Objective-C automatically keeps track of this for you.

Currently, our Fraction class declares (and later implements) three methods. All are instance methods, as noted by the minus sign () preceding each method name. A class method is delineated by a plus sign (+).

Before we discuss the method declaration, we must remember than an object is considered a “black box.” Ideally, objects encapsulate, or hide, their instance variables, from anything except itself. The object exists as a single entity, and anything that works with the object does not need to know, and should not know, about the object’s inner workings. Therefore, we need to declare specific methods to set the variables, because the program should not directly manipulate them. We also need a separate method to print the values, because, once again, the program should not be able to directly access the variables. This is a fundamental concept in Objective-C, and is observed throughout the libraries and code.

The Fraction class defines three methods, all of which return nothing. The value in parentheses before the method name specifies the return type of the method. void is programming parlance for “nothing,” or nil. Any data type or object can be a return value; a method might return a float, int, or another Fraction, as illustrated by the following methods:

- (int)numerator;
- (float)divide;
- (Fraction *)randomFraction;

The asterisk after the Fraction is used to delineate a pointer; this is not an important concept at the moment. For now, simply remember that anytime you create an instance of a custom object, the name of the instance should be prefixed with an asterisk.

All of these methods must return a value somewhere in the method. Just as main returns 0 at the end, these methods must return an int value, a float value, and another instance of Fraction, respectively. The compiler will complain if you do not return a value.

The latter two methods declared in Fraction return no value, but they do each take an integer value as an argument. In the case of setNumerator:, the name of the argument is n. This name is arbitrary, as long as it follows the standard naming conventions, and is used only within the method to refer to the argument. The argument itself exists only within the method; you cannot access n anywhere else in Fraction.

Arguments are values that are passed into a method, which the method would not be able to obtain otherwise. For example, setNumerator: requires a value to set to the numerator, but it would not receive a custom value without an argument.

Note the syntax of these method declarations. The method name begins with the minus or plus, signifying the method type. This is followed by the return type in parentheses, the name of the method, and a colon. This colon is followed by an argument type in parentheses, a name for the argument, and a semicolon. In a method that takes an argument, the colon is considered part of the method name.

A method can take multiple arguments; they are delineated by a space, a new section of the method name, another colon, an argument type, and name. This will be covered in a later chapter.

The @implementation Section

In the implementation section, you define the code that each method implements. The general appearance of the @implementation section is similar to this:

@implementation ClassName

In the implementation of Fraction, you also declare the method types and names, followed by an open brace. Between the opening and closing brace, you insert the code that the method will execute when it is called.

The program Section

In this program, the main program section contains the all-important main function that is the first thing to be executed.

In Program 3.2, we first create an instance of Fraction:

 Fraction *myFraction;

Again, don’t worry about the asterisk. From a technical perspective, it says that myFraction is a pointer to a Fraction.

Now that you’ve created an object “container” to store a Fraction, you have to actually create a Fraction to fill that space. You do that in the next line,

myFraction = [[Fraction alloc] init];

In this line, you call the alloc method on Fraction; alloc is a class method. Although you never defined the alloc method, it is inherited from NSObject. Inheritance will be discussed in a later lesson.

After you allocate a Fraction, you have to actually initialize it. Therefore, you pass the value of the alloc method to NSObject’s init method, and the resulting Fraction is assigned to myFraction. This nested sequence of messages is used almost every time you create a new object.

Next, you set the values of myFraction to 2 and 5 by using the proper method calls.

[myFraction setNumerator:2];
[myFraction setDenominator:5];

You pass the integers 2 and 5 as arguments to the method. The method then takes these arguments and sets the numerator and denominator variables of myFraction.

You then call myFraction’s showResults method to print the value of the fraction.

We then free up the memory that Fraction uses by calling the release method:

[myFraction release];

This is a critical part of programming for iOS. Whenever you create a new object through alloc/init, you have to release it—in other words, every alloc/init call is balanced by a release call. Otherwise, the now-useless object will still sit in memory, taking up space; in a memory-constrained environment such as the iPhone, you want to use as little memory as you can.

Messaging Receivers

In every message, you send the message to a specific receiver—the object or class which implements the method. However, instance methods are sent to specific instances, so that the same method call to two different Fractions will result in two Fractions with potentially different values. For example,

[fraction1 setNumerator:2];
[fraction1 setDenominator:5];

[fraction2 setNumerator:3];
[fraction2 setDenominator:4];

Will result in two separate fractions, one with a value of 2/5 and another with a value of 3/4. Although you are calling the same methods on both, they are received by different objects, with their own set of numerator and denominator variables.

Data Encapsulation

As mentioned before, you shouldn’t directly access the instance variables of an object. Data encapsulation provides some security against people working with the class, to prevent them from tinkering with the inner workings. So in the main function, how would you access these values?

Generally, you access these values by writing special methods, known as getter methods (the set methods that we’ve already wrote for Fraction are known as setter methods; they are generally written in pairs).

The getter methods for the Fraction class look something like this:

- (int)numerator;
- (int)denominator;

- (int)numerator {
	return numerator;
- (int)denominator {
	return denominator;

It doesn’t matter if the method name is the same as the variable name; in fact, it is common practice.

Program 3.3 contains the main function that tests these new methods.

Program 3.3

int main (int argc, const char * argv[]) {
    NSAutoreleasePool * pool = [[NSAutoreleasePool alloc] init];

    // Create an instance of Fraction
	Fraction *myFraction;

	myFraction = [[Fraction alloc] init];

	// Set myFraction to 2/5
	[myFraction setNumerator:2];
	[myFraction setDenominator:5];

	// Display the value of myFraction
	NSLog (@"The value of myFraction is %d/%d", [myFraction numerator], [myFraction denominator]);

	// Clear up memory
	[myFraction release];    [pool drain];
    return 0;

The output is

The value of myFraction is 2/5

Objective-C 2.0 introduces properties, which provide an easy way to create getter and setter methods. This will also be covered in a later lesson.

  • Welcome

    My goal is to make CupsOfCocoa into a beautiful source for beginners to the iPhone platform to get started. Subscribe below for more, and stay tuned!

  • Contact Me

    If you need to contact me for any reason, feel free to send me an email.
  • The Giving Spirit

    If you've found this site helpful, would you consider donating a little sum? Any amount is appreciated...Thanks so much!

  • Roadmap

  • Enter your email address to follow this blog and receive notifications of new posts by email.

    Join 220 other followers

  • Back to the Past

    May 2020
    S M T W T F S
  • Time Machine

  • You count!

  • Worldwide Stats

    free counters
%d bloggers like this: