Latest update on Thursday 16 October 2008 à 09:43 by Jean-François Pillou.
Definition of polymorphism
The word polymorphism comes from Greek and means having several different forms.
This is one of the essential concepts of object-oriented programming. Where inheritance is related to classes and
(their hierarchy), polymorphism is related to object methods.
In general there are three types of polymorphism:
Overloading polymorphism
Parametric polymorphism (also called template polymorphism)
Inclusion polymorphism (also called redefinition or overriding)
We will now attempt to describe more precisely the above types, but be warned, many people get
confused when trying to understand the differences between the three types.
Overloading polymorphism
Overloading polymorphism is where functions of the same name exist, with similar functionality,
in classes which are completely independent of each other (these do not have to be children of the object class).
For example, the complex class, the image class and the link class may each have the function "display".
This means that we do not need to worry about what type of object we are dealing with if all we want to do is
display it on the screen.
Overloading polymorphism therefore allows us to define operators whose behavior will vary
depending on the parameters that are applied to them. Therefore it is possible, for example, to add
the + operator and make it behave differently according to whether it refers to an operation between two
integers (addition) or between two character strings (concatenation).
Parametric polymorphism
Parametric polymorphism is the ability to define several functions using the same name, but
using different parameters (name and/or type). Parametric polymorphism automatically selects the correct
method to be adopted according to the type of data passed in the parameter.
We can therefore define several addition() methods using the same name which
calculates a sum of values.
The int addition(int,int) method would return the sum of two integers.
The float addition(float, float) would return the sum of two floats.
The char addition(char, char) would result as the sum of two characters as defined by the author.
etc.
A signature is the name and type (static) given to the arguments of a function. Therefore
it is a method's signature which determines what is called on.
Inclusion polymorphism
The ability to redefine a method in classes that are inherited from a base class is
called specialization. One can therefore call on an object's method without having to know its intrinsic
type: this is inclusion polymorphism. This makes it possible to disregard the specialized class details of an
object family, by masking them with a common interface (this being the basic class).
Imagine a game of chess, with the objects king, queen, bishop, knight,
rook and pawn, each inheriting the piece object. The movement method could, using
inclusion polymorphism, make the corresponding move according to the object class that is called on. This therefore
allows the program to perform piece.movement without having to be concerned with each piece's class.
This document, titled « OOP - Polymorphism », is available under the Creative Commons license. Any copy, reuse, or modification of the content should be sufficiently credited to CCM (ccm.net).
