A Primer on Object-Oriented Concepts
Like this article? We recommend
Like this article? We recommend
Like this article? We recommend
If you're like me, early on in the OO language world, you didn't hear much about OO concepts and how they apply to good application design. It may be why nearly all early, large OO applications developed in the late ‘90s to early 2000s were poorly designed, even if using any design conventions at all. If these apps haven't been “reworked” to utilize modern-day web-app concepts, they are difficult to work with and hard to maintain, which means just keeping things at the status quo. A well-designed app can grow in usage and is always easy to use and easy to maintain and extend. For businesses creating web-based applications, this means increased efficiency and getting faster to market with new features. Overall, it means saving money and growing your business.
In this article, you see how OO concepts apply to good application design. If you are new to these concepts, hopefully you can appreciate how effective they are in understanding how these concepts make OO languages more popular over procedural languages. If you are already familiar with these concepts, maybe there will be some new things you didn't know about them.
Core OO Concepts
Encapsulation
The idea behind this concept is that your OO classes are essentially a black box. Users of the class should not know how it works internally and neither should other classes. An example would be using a Calculator class. A user simply types in the equation and then gets the answer. How the calculator arrived at the answer is hidden from the user. (Although the user in this case likely has a good idea.) Also important is that other classes that use the Calculator class do not know how the answer was obtained. The calculator's internal logic is not visible and as such, the class is encapsulated.
To encapsulate functionality of a class in an OO language, an Interface class is used. You can think of an Interface class as the declarations to a set of methods in a class. The Interface is all the user and other classes have access to. The actual implementation of the class is hidden. For example, the interface to a Calculator class could be
add(X, Y) (returns a String) subtract (X, Y) (returns a String) divide(X,Y) (returns a String) multiply(X,Y) (returns a String)
To use the interface, another method simply calls the operation with some numbers, that is, add(4,5). The answer is returned as a string to the class that invoked the interface:
ICalculator calculator = new ICalculator(); String result = calculator.add(4,5);
Something else the interface does is enable the functionality of a class to be changed without having to change this code anywhere else. The methods that use the interface do not need to be changed in any way. This is great for testing with different implementations or changing and extending functionality.
Another good reason for using interfaces is that they are contracts on how a method should be implemented. By specifying what method declarations can be used in an interface, this determines how the method should be coded. A good example of interfaces acting as contracts is the Java specification. Java specifications (that is, JPAs) define a contract as to what methods can be coded and how (what to pass in as variables, and so on). Interfaces are an important part of many popular design patterns.
Are there any disadvantages to using interfaces? Yes, but very few. A disadvantage to using an interface is that users of the interface must implement all methods defined in the interface. Although this enforces the contract part of the interface, many methods an interface defines are not necessary. For example, large business applications often have large interfaces used by all clients; although, only some of the operations apply or are relevant. This leads you to the Interface Segregation Principal. The principal states that any interfaces that are large and do not apply to all clients should be broken down into smaller interfaces. Breaking large interfaces down to smaller interfaces ensures that only some interfaces will be used and not others, depending on their relevance for users of the interface. These smaller interfaces are often referred to as role interfaces.
Inheritance
Probably the most discussed OO concept is Inheritance. Several design patterns also use inheritance. The concept of Inheritance is that one class inherits the methods of another class. Often, the classes inherited are a parent class of an object. For example, a Circle class would inherit the parent class methods of a class or interface called Shape. Circle would then override the methods defined in Shape. In Java, the code to inherit an interface would look like
class Circle implements Shape
If Shape is an interface, then other objects who share the same attributes (that is, color, height, and width) can also use Shape. For example, Square could also implement (inherit) the attributes Shape provides. The advantage to inheritance is that you can abstract out common attributes that are similar to a set of objects. In this example, the Shape class has methods and attributes that other objects need to implement, along with their own methods. A Circle would implement method operations and attributes that are exclusive only to a circle (that is, radius), along with those inherited from Shape. Can a class inherit multiple other classes? Yes, though in Java, you can do so only with interfaces and abstract classes. With Java, by extending multiple interfaces, you are essentially doing the same thing as inheriting from multiple classes. The caveat, though, is that with interfaces, you are required to implement all method declarations of said interfaces. With abstract classes, however, you do not have to implement all methods as with interfaces. You can think of abstract classes as partial classes. The advantage to inheriting from abstract classes is that you do not have to implement/override all methods of the abstract class.
There are three ways for subclasses to inherit and override/implement methods from an abstract (parent) class:
- If a base class method is abstract, the subclass can override this method.
- If a base class method has a constructor with a concrete implementation, a subclass must override this method of the base class.
- If a base class has a public, static, and final method, no subclass can (or should) override this method of this base class.
Composition
Before wrapping up inheritance, you should also know there are basically two ways a subclass can inherit from a parent class. Composition is the term used to describe the relationship between the parent and child objects (or base and subclass). There are two types of compositions: association and aggregation. An aggregation composition is an object composed of other objects forming a complex component. An example would be a car. A car has an engine, pistons, and so on. The relationship between the car and its parts is an aggregation. An association composition is a relationship that defines a service for the child object provided by the parent object. For example, a car has a garage. The garage is the service component because it complements the car but is not part of a car.
Polymorphism
Polymorphism means that an interface or abstract class has the capacity to take on different forms by representing different objects when accessed by different methods. A good example of polymorphism in Java is your factory classes. A factory class returns different object types based on what was passed into the factory from a calling method. A simple example of this would be an abstract class called Car acting as the base class used by a factory class:
public abstract class Car{
public abstract String make();
}
Some subclasses of Car could be Oldsmobile and Tesla:
public class Oldsmobile extends Car {
@Override
public String make() {
return "Oldsmobile"
}
}
public class Tesla extends Car {
@Override
public String make() {
return "Tesla"
}
}
You can get different responses using the same abstract class to determine the vehicle make when passing in an attribute specific for that make to a factory class:
public class CarFactory {
public Car getCar(String type) {
if ("electric".equals(type)) {
return new Tesla();
if ("cutless".equals(type)) {
return new Oldsmobile();
}
}
}
Testing this factory with a driver class, you have
public class Demo {
public static void main(String[] args) {
CarFactory carFactory = new CarFactory();
Car c1 = carFactory.getCar("electric");
System.out.println("c1 Make: " + c1.make());
Car c2 = carFactory.getCar("cutless");
System.out.println("c2 Make: " + c2.make());
}
}
By using the same abstract class and returning different types, the definition for polymorphism is supported by a factory class. You could easily replace the abstract class with an interface.
Conclusion
This article was a primer for those who may need a refresher on OO concepts to aid in better application design. By revisiting or learning these concepts for the first time, you can benefit by providing more robust applications while reducing maintenance. You also learned how factory classes can be good examples of polymorphism.