Object-Oriented Programming Concepts and .NET
Part 2: Inheritance, Abstraction, and Polimorphism in .NET
Erika Ehrli Cabral
March 2005
Summary
The following article is the second of a three-part article series that presents
definitions and samples for different Object-Oriented
Programming (OOP) concepts and its
implementation in .NET. The first part examined the concepts of classes, objects,
and structures. This part examines the concepts of inheritance, abstraction, and polimorphism.
The third and last part will examine the concepts of interface, multiple interface
inheritance, collections, and overloading.
Contents
Introduction
In Part 1 of Object-Oriented Programming Concepts and .NET, I defined the concepts
of class, object, and structure. In addition to defining the concepts, I explained
real world samples and presented sample code in C# and VB.NET to create classes
and structs. The first article also explains objects as independent building blocks.
In Part 2 of Object-Oriented Programming Concepts and .NET, I will explain the concepts
of inheritance, abstraction, and polimorphism. I will also present a Unified Model
Language (UML) class diagram to represent an object model that will help as a visual
aid to explain some concepts. The purpose of this article is to explain a series of
relationships between objects.
Inheritance
In the real world there are many objects that can be specialized. In OOP, a parent
class can inherit its behavior and state to children classes. This concept was developed
to manage generalization and specialization in OOP and is represented by a is-a relationship.
The following OO terms are commonly used names given to parent and child classes
in OOP:
- Superclass: Parent class.
- Subclass: Child class.
- Base class: Parent class.
- Derived class: Child class
The most common real world sample to explain inheritance is the geometric shapes
object model. Squares, circles, triangles,
rectangles, pentagons, hexagons, and
octagons are geometric shapes.
The following figure shows a sample set of geometric figures:
Figure 1. Geometric shapes.
The concept of generalization in OOP means that an object encapsulates
common state an behavior for a category of objects. The general object in this sample
is the geometric shape. Most geometric shapes
have area, perimeter, and color. The
concept of specialization in OOP means that an object can inherit the common
state and behavior of a generic object; however, each object needs to define its
own special and particular state an behavior. In Figure 1, each shape has its own
color. Each shape has also particular formulas to calculate its area and perimeter.
Inheritance makes code elegant and less repetitive. If we know that all shapes have
color, should we program a color attribute for each shape? The answer is no! Would
it be a better idea to create a shape class that has a color attribute and to make
all the specialized shapes to inherit the color attribute? The answer is yes!
An object model for this sample could have a shape parent
class and a derived class for each specific shape.
The following UML class diagram shows the set of classes needed to
model the geometric shapes sample. Observe the field, properties, and methods for
each class:
Figure 2. The Shape class is the parent class. Square, Rectangle, and Circle
are derived classes that inherit from Shape. The triangle-connector in
the diagram represents an is-a relationship.
The .NET framework has many base classes. Everything is derived from System.Object.
You can create almost anything you imagine using the built-in functionality provided
in the
.NET Framework Class Library.
To create a derived class in C#, the class declaration should be done as:
class child: parent
To create a derived class in VB.NET, the class declaration should be done as:
Class child
Inherits parent
End Class
Multiple inheritance
Multiple inheritance is the possibility that a child class can have multiple parents.
Human beings have always two parents, so a child will have characteristics from
both parents.
In OOP, multiple inheritance might become difficult to handle because it allows
ambiguity for the compiler. There are programming languages such as C++ that allow
multiple inheritance; however, other programming languages such as Java and the
.NET Framework languages do not allow multiple inheritance. Multiple inheritance
can be emulated in .NET using Multiple Interface Inheritance, which I will explain
in Part 3 of this series.
Sealed class
A sealed class is a class that does not allow inheritance. Some object model
designs need to allow the creation of new instances but not inheritance, if this is the case, the class
should be declared as sealed.
To create a sealed class in C#, the class declaration should be done as:
sealed class Shape
To create a sealed class in VB.NET, the class declaration should be done as:
NotInheritable Class Shape
Abstraction
Abstraction is "the process of identifying common patterns that have systematic
variations; an abstraction represents the common pattern and provides a means for
specifying which variation to use" (Richard Gabriel).
An abstract class is a parent class that allows inheritance but can never be instantiated.
Abstract classes contain one or more abstract methods that do not have implementation.
Abstract classes allow specialization
of inherited classes.
Figure 2 shows a Shape class, which is an abstract class. In the real world,
you never calculate the area or perimeter of a generic shape, you must know what
kind of geometric shape you have because each shape (eg. square, circle, rectangle,
etc.) has its own area and perimeter formulas. The parent class shape forces all
derived classes to define the behavior for CalculateArea() and CalculatePerimeter().
Another great example is a bank account. People own savings accounts, checking accounts,
credit accounts, investment accounts, but not generic bank accounts. In this case,
a bank account can be an abstract class and all the other specialized bank accounts
inherit from bank account.
To create an abstract class in C#, the class declaration should be done as:
abstract class Shape
To create an abstract class in VB.NET, the class declaration should be done as:
MustInherit Class Shape
To following code shows a sample implementation of an abstract class:
/// C#
using System;
namespace DotNetTreats.OOSE.OOPSamples{
public abstract class Shape{
private float _area;
private System.Drawing.Color _color;
private float _perimeter;
public float Area{
get{
return _area;
}
set{
_area = value;
}
}
public System.Drawing.Color Color{
get{
return _color;
}
set{
_color = value;
}
}
public float Perimeter{
get{
return _perimeter;
}
set{
_perimeter = value;
}
}
public abstract void CalculateArea();
public abstract void CalculatePerimeter();
}
}
Listing 1. The Shape abstract class in C#.
Polimorphism
Polimorphism allows objects to be represented in multiple forms. Even though classes
are derived or inherited from the same parent class, each derived class will have
its own behavior. Polimorphism is a concept linked to inheritance and assures
that derived classes have the same functions even though each derived class performs
different operations.
Figure 2 shows a Rectangle, a Circle, and Square. All of them are shapes and
as shapes their area and perimeter can be calculated; however, each shape calculates its
area in a specialized way. Declaring a member as abstract allows polimorphism.
The Shape class defines the CalculateArea() and CalculatePerimeter()
methods as abstract, this allows each derived class to override the implementation
of the parent's methods.
To following sample code shows an implementation of a derived class (rectangle).
The specific CalculateArea() and CalculatePerimeter() methods
for the rectangle class illustrate polimorphism:
/// C#
using System;
namespace DotNetTreats.OOSE.OOPSamples{
class Rectangle
: Shape{
private float _height;
private float _width;
public rectangle(float height, float width){
_height = height;
_width = width;
}
public float Height{
get{
return _height;
}
set{
_height = value;
}
}
public float Width{
get{
return _width;
}
set{
_width = value;
}
}
public override void CalculateArea(){
this.Area = _height * _width;
}
public override void CalculatePerimeter(){
this.Perimeter = (_height * 2) + (_width * 2);
}
}
}
Listing 2. Polimorphism represented in the Rectangle's methods.
Virtual keyword
The virtual keyword allows polimorphism too. A virtual property or method has an implementation in the base class, and can be overriden in the derived classes.
To create a virtual member in C#, use the virtual keyword:
public virtual void Draw()
To create a virtual member in VB.NET, use the Overridable keyword:
Public Overridable Function Draw()
Override keyword
Overriding is the action of modifying or replacing the implementation of the parent
class with a new one. Parent classes with virtual or abstract members allow derived
classes to override them.
To override a member in C#, use the override keyword:
public override void CalculateArea()
To override a member in VB.NET, use the Overrides keyword:
Public Overrides Function CalculateArea()
Conclusion
Inheritance allows developers to manage a generalization and specialization relationship
between objects. OOP concepts such as abstraction and polimorphism help to define
better object models where object hierarchies are designed with reusability in mind.
In this article, I examined the concept of inheritance, abstraction, and polimorphism.
The third and last part of this series will examine the concepts of interface, multiple
interface inheritance, collections, and overloading.
Note: The sample source code*
for this article works only in Visual Studio 2005.
Reference
- Matt Weisfeld, The Object-Oriented Thought Process,
SAMS, 2000.
- Robin A. Reynolds-Haertle, OOP with Microsoft Visual
Basic .NET and Microsoft Visual C# .NET Step by Step, Microsoft Press, 2002.