Polymorph Is M

Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

CHAPTER 1 N PROGRAMMING CONCEPTS

Polymorphism
The word “polymorphism” has its root in two Greek words: “poly” (means many) and “morphos” (means form).
In programming, polymorphism is the ability of an entity (e.g. variable, class, method, object, code, parameter,
etc.) to take on different meanings in different contexts. The entity that takes on different meanings is known as
a polymorphic entity. Various types of polymorphism exist. Each type of polymorphism has a name that usually
indicates how that type of polymorphism is achieved in practice. The proper use of polymorphism results in generic
and reusable code. The purpose of polymorphism is writing reusable and maintainable code by writing code in terms
of a generic type that works for many types (or ideally all types).
Polymorphism can be categorized in the following two categories:
u Ad hoc polymorphism
u Universal polymorphism
If the types for which a piece of code works are finite and all those types must be known when the code is written,
it is known as ad hoc polymorphism. Ad hoc polymorphism is also known as apparent polymorphism because it
is not a polymorphism in a true sense. Some computer science purists do not consider ad hoc polymorphism as
polymorphism at all. Ad hoc polymorphism is divided into two types: overloading polymorphism and coercion
polymorphism.
If a piece of code is written in such a way that it works for infinite number of types (will also work for new types
not known at the time the code is written), it is called universal polymorphism. In universal polymorphism, the same
code works on many types, whereas in ad hoc polymorphism, different implementations of code are provided for
different types giving an apparent impression of polymorphism. Universal polymorphism is divided into two types:
inclusion polymorphism and parametric polymorphism.

Overloading Polymorphism
Overloading is an ad hoc polymorphism. Overloading results when a method (called a method in Java and a function
in other languages) or an operator has at least two definitions that work on different types. In such cases, the same
method or operator name is used for their different definitions. That is, the same name exhibits many behaviors and
hence the polymorphism. Such methods and operators are called overloaded methods and overloaded operators.
Java lets you define overloaded methods. Java has some overloaded operators. Java does not let you overload an
operator for an ADT. That is, you cannot provide a new definition for an operator in Java.
Listing 1-7 shows code for a class named MathUtil.

Listing 1-7. An Example of an Overloaded Method in Java


package com.jdojo.concepts;

public class MathUtil {


public static int max(int n1, int n2) {
/* Code to determine the maximum of two integers goes here */
}

public static double max(double n1, double n2) {


/* Code to determine the maximum of two floating-point numbers goes here */
}

public static int max(int[] num) {


/* Code to determine the maximum of an array of int goes here */
}
}

25
www.acm.org.ir
CHAPTER 1 N PROGRAMMING CONCEPTS

The max() method of the MathUtil class is overloaded. It has three definitions and each of its definitions
performs the same task of computing maximum, but on different types. The first definition computes a maximum of
two numbers of int data type, the second one computes a maximum of two floating-point numbers of double data
type, and the third one computes a maximum of an array of numbers of int data type. The following snippet of code
makes use of all three definitions of the overloaded max() method:

int max1 = MathUtil.max(10, 23); // Uses max(int, int)


double max2 = MathUtil.max(10.34, 2.89); // Uses max(double, double)
int max3 = MathUtil.max(new int[]{1, 89, 8, 3}); // Uses max(int[])

Note that method overloading gives you only sharing of the method name. It does not result in the sharing of
definitions. In Listing 1-7, the method name max is shared by all three methods, but they all have their own definition
of computing maximum of different types. In method overloading, the definitions of methods do not have to be
related at all. They may perform entirely different things and share the same name.
The following code snippet shows an example of operator overloading in Java. The operator is +. In the following
three statements, it performs three different things:

int n1 = 10 + 20; // Adds two integers


double n2 = 10.20 + 2.18; // Adds two floating-point numbers
String str = "Hi " + "there"; // Concatenates two strings

In the first statement, the + operator performs addition on two integers, 10 and 20, and returns 30. In the second
statement, it performs addition on two floating-point numbers, 10.20 and 2.18, and returns 12.38. In the third
statement, it performs concatenation of two strings and returns "Hi there".
In overloading, the types of the actual method’s parameters (types of operands in case of operators) are used
to determine which definition of the code to use. Method overloading provides only the reuse of the method name.
You can remove method overloading by simply supplying a unique name to all versions of an overloaded method.
For example, you could rename the three versions of the max() method as max2Int(), max2Double(), and maxNInt().
Note that all versions of an overloaded method or operator do not have to perform related or similar tasks. In Java, the
only requirement to overload a method name is that all versions of the method must differ in number and/or type of
their formal parameters.

Coercion Polymorphism
Coercion is an ad hoc polymorphism. Coercion occurs when a type is implicitly converted (coerced) to another type
automatically even if it was not intended explicitly. Consider the following statements in Java:

int num = 707;


double d1 = (double)num; // Explicit conversion of int to double
double d2 = num; // Implicit conversion of int to double (coercion)

The variable num has been declared to be of int data type, and it has been assigned a value of 707. The second
statement uses cast, (double), to convert the int value stored in num to double, and assigns the converted value
to d1. This is the case of explicit conversion from int to double. In this case, the programmer makes his intention
explicit by using the cast. The third statement has exactly the same effect as the second one. However, it relies on
implicit conversion (called widening conversion in Java) provided by Java language that converts an int to double
automatically when needed. The third statement is an example of coercion. A programming language (including Java)
provides different coercions in different contexts: assignment (shown above), method parameters, etc.

26
www.acm.org.ir
CHAPTER 1 N PROGRAMMING CONCEPTS

Consider the following snippet of code that shows a definition of a square() method, which accepts a parameter
of double data type:

double square(double num) {


return num * num;
}

The square() method can be called with actual parameter of double data type as

double d1 = 20.23;
double result = square(d1);

The same square() method may also be called with actual parameter of int data type as

int k = 20;
double result = square(k);

You have just seen that the square() method works on double data type as well as int data type although you
have defined it only once in terms of a formal parameter of double data type. This is exactly what polymorphism
means. In this case, the square() method is called a polymorphic method with respect to double and int data type.
Thus the square() method is exhibiting polymorphic behavior even though the programmer who wrote the code
did not intend it. The square() method is polymorphic because of the implicit type conversion (coercion from int to
double) provided by Java language. Here is a more formal definition of a polymorphic method:

Suppose m is a method that declares a formal parameter of type T. If S is a type that can be implicitly
converted to T, the method m is said to be polymorphic with respect to S and T.

Inclusion Polymorphism
Inclusion is a universal polymorphism. It is also known as subtype (or subclass) polymorphism because it is achieved
using subtyping or subclassing. This is the most common type of polymorphism supported by object-oriented
programming languages. Java supports it. Inclusion polymorphism occurs when a piece of code that is written using
a type works for all its subtypes. This type of polymorphism is possible based on the subtyping rule that a value that
belongs to a subtype also belongs to the supertype. Suppose T is a type and S1, S2, S3... are subtypes of T. A value that
belongs to S1, S2, S3... also belongs to T. This subtyping rule makes us write code as follows:

T t;
S1 s1;
S2 s2;
...
t = s1; // A value of type s1 can be assigned to variable of type T
t = s2; // A value of type s2 can be assigned to variable of type T

27
www.acm.org.ir
CHAPTER 1 N PROGRAMMING CONCEPTS

Java supports inclusion polymorphism using inheritance, which is a subclassing mechanism. You can define
a method in Java using a formal parameter of a type, for example, Person, and that method can be called on all its
subtypes, for example, Employee, Student, Customer, etc. Suppose you have a method processDetails() as follows:

void processDetails(Person p) {
/* Write code using the formal parameter p, which is of type Person. The same code will
work if an object of any of the subclass of Person is passed to this method.
*/
}

The processDetails() method declares a formal parameter of Person type. You can define any number of
classes that are subclasses of the Person class. The processDetails() method will work for all subclasses of the
Person class. Assume that Employee and Customer are subclasses of the Person class. You can write code like

Person p1 = create a Person object;


Employee e1 = create an Employee object;
Customer c1 = create a Customer object;
processDetails(p1); // Use Person type
processDetails(e1); // Use Employee type, which is a subclass of Person
processDetails(c1); // Use Customer type, which is a subclass of Person

The effect of the subtyping rule is that the supertype includes (hence the name inclusion) all values that belong
to its subtypes. A piece of code is called universally polymorphic only if it works on an infinite number of types.
In the case of inclusion polymorphism, the number of types for which the code works is constrained but infinite.
The constraint is that all types must be the subtype of the type in whose term the code is written. If there is no
restriction on how many subtypes a type can have, the number of subtypes is infinite (at least in theory). Note that
inclusion polymorphism not only lets you write reusable code, it also lets you write extensible and flexible code. The
processDetails() method works on all subclasses of the Person class. It will keep working for all subclasses of the
Person class, which will be defined in future, without any modifications. Java uses other mechanisms, like method
overriding and dynamic dispatch (also called late binding), along with subclassing rules to make the inclusion
polymorphism more effective and useful.

Parametric Polymorphism
Parametric is a universal polymorphism. It is also called “true” polymorphism because it lets you write true generic
code that works for any types (related or unrelated). Sometimes it is also referred to as generics. In parametric
polymorphism, a piece of code is written in such a way that it works on any type. Contrast parametric polymorphism
with inclusion polymorphism. In inclusion polymorphism, code is written for one type and it works for all of its
subtypes. It means all types for which the code works in inclusion polymorphism are related by a supertype-subtype
relationship. However, in parametric polymorphism, the same code works for all types, which are not necessarily
related. Parametric polymorphism is achieved by using a type variable when writing the code, rather than using any
specific type. The type variable assumes a type for which the code needs to be executed. Java supports parametric
polymorphism since Java 5 through generics. Java supports polymorphic entity (e.g. parameterized classes) as well as
polymorphic method (parameterized methods) that use parametric polymorphism.
All collection classes as of Java 5 have been retrofitted to use generics (parametric polymorphism is achieved
in Java using generics). You can write code using generics as shown below. This code uses a List object as a list of
String type and a List object as a list of Integer type. Using generics, you can treat a List object as a list of any type
in Java. Note the use of <XXX> (angle brackets) in code to specify the type for which you want to instantiate the
List object.

28
www.acm.org.ir
CHAPTER 1 N PROGRAMMING CONCEPTS

// Use List for String type


List<String> sList = new ArrayList<String>();
sList.add("string 1");
sList.add("string 2");
String s2 = sList.get(1);

// Use List for Integer type


List<Integer> iList = new ArrayList<Integer>();
iList.add(10);
iList.add(20);
int i2 = iList.get(1);

Summary
Writing the set of instructions for a computer to accomplish a task is known as programming. The set of instructions is
known as a program. Different types of programming languages exist. They differ in their closeness to the instructions
that the hardware can understand or the paradigm. A machine language lets you write programs using 0s and 1s,
and it is the lowest level programming language. A program written in machine language is known as machine code.
An assembly language lets you write programs using mnemonics. A program written using an assembly language is
known as assembly code. Later, higher-level programming languages were developed, using an English-like language.
Several types of programming paradigms are in practice. A programming paradigm is a thinking cap for viewing
and analyzing real-world problems in a particular way. Imperative, procedural, functional, and object-oriented are
some widely used paradigms in software development. Java is a programming language that supports procedural,
functional, and object-oriented programming paradigms.
Abstraction, encapsulation, inheritance, and polymorphism are the four pillars of object-oriented paradigms.
Abstraction is the process of hiding details of a program that are irrelevant to the users. Encapsulation is the process
of bundling multiple items into one entity. Inheritance is the process of arranging classes in a hierarchical manner to
build supertype/subtype relationships. Inheritance promotes reusability of code by allowing programmers to write
the code in terms of a supertype that also works for all of the subtypes. Polymorphism is the way of writing a piece of
code once that can operate on multiple types. Method overloading, method overriding, subtyping, and generics are
some of the ways to implement polymorphism.

29
www.acm.org.ir

You might also like