Use Abstract class or Interface?

When starting out, it is difficult to understand when to use an abstract class or an interface. This blog is mean to help you understand it.

In programming, it is really important to implement good coding practices like OOP inheritance, using abstract classes and interfaces. Here is a very brief summary on both topics. Further information can be found on internet.

Abstract class

It declares fields, properties and methods that can be used for any class that inherits from it. The method implementation can be abstract (mandatory) or virtual (optional)

Interfaces

It declares methods, properties and events.
It is a contract, so each class that inherits from it must implement its declared members.

So, when to use Abstract class or Interface?

Abstract classes: when creating several objects of the same type

For example: soccerball, cricketball and basketball are all balls. So it makes sense to create a base class called ball

Interfaces: several types of objects behave the same way

A player and an enemy can move, shoot and take damage, so they can implementIMove, IShoot and IDamage

Implementing this simple UML diagram

The pics: read from top to bottom and then vice versa

The process:

DamageOnClick

• When click on screen, if hit onCar1 or Barrel1, it will try to execute TakeDamage(1)

Case 1: Barrel1

• DamageOnClick will look for ITakeDamage in Barrel1 script
• As it inherits from BaseDamage and it inherits ITakeDamage as well, the condition if(iDamage != null) in DamageOnClickis true
• TakeDamage() will be executed with a value of 1 in accordance with parent’s ITakeDamage implementation
• The maxHealth value must be set in inspector

Case 2: Car1

• First 3 steps as previously
• TakeDamage() will be executed with a value of 3 because Car1 overrides its parent’s value.
• The maxHealth is set with hardcode when overriding SetMaxHealth()

Why are abstract classes and interfaces important?

Without interfaces, when click on the GameObject, the code will look some like this:

The developer would have to evaluate each possible GameObject that could be hit when click on screen.

In this case there are only 2 GameObjects, but imagine if they would be 30 or 100 hundred types of enemies…

On the other hand, when using interfaces

The code is shorter, cleaner, and most importantly, the if condition will check the implementation of TakeDamage() on the object that has been hit through the ITakeDamage implementation. That’s all.

The developer doesn’t need to explicitly code every possible case where a GameObject can be hit.

For example, if I have 20 different types of enemies, each one must to implement the TakeDamage() method, and those 2 lines of code will be enough to execute it, no matter the type of enemy.