Interface vs Abstract Class (general OO)

While your question indicates it's for "general OO", it really seems to be focusing on .NET use of these terms.

In .NET (similar for Java):

• interfaces can have no state or implementation
• a class that implements an interface must provide an implementation of all the methods of that interface
• abstract classes may contain state (data members) and/or implementation (methods)
• abstract classes can be inherited without implementing the abstract methods (though such a derived class is abstract itself)
• interfaces may be multiple-inherited, abstract classes may not (this is probably the key concrete reason for interfaces to exist separately from abtract classes - they permit an implementation of multiple inheritance that removes many of the problems of general MI).

As general OO terms, the differences are not necessarily well-defined. For example, there are C++ programmers who may hold similar rigid definitions (interfaces are a strict subset of abstract classes that cannot contain implementation), while some may say that an abstract class with some default implementations is still an interface or that a non-abstract class can still define an interface.

Indeed, there is a C++ idiom called the Non-Virtual Interface (NVI) where the public methods are non-virtual methods that 'thunk' to private virtual methods:

• http://www.gotw.ca/publications/mill18.htm
• http://en.wikibooks.org/wiki/More_C%2B%2B_Idioms/Non-Virtual_Interface

Short: Abstract classes are used for Modelling a class hierarchy of similar looking classes (For example Animal can be abstract class and Human , Lion, Tiger can be concrete derived classes)

AND

Interface is used for Communication between 2 similar / non similar classes which does not care about type of the class implementing Interface(e.g. Height can be interface property and it can be implemented by Human , Building , Tree. It does not matter if you can eat , you can swim you can die or anything.. it matters only a thing that you need to have Height (implementation in you class) ).

There are a couple of other differences -

Interfaces can't have any concrete implementations. Abstract base classes can. This allows you to provide concrete implementations there. This can allow an abstract base class to actually provide a more rigorous contract, wheras an interface really only describes how a class is used. (The abstract base class can have non-virtual members defining the behavior, which gives more control to the base class author.)

More than one interface can be implemented on a class. A class can only derive from a single abstract base class. This allows for polymorphic hierarchy using interfaces, but not abstract base classes. This also allows for a pseudo-multi-inheritance using interfaces.

Abstract base classes can be modified in v2+ without breaking the API. Changes to interfaces are breaking changes.

[C#/.NET Specific] Interfaces, unlike abstract base classes, can be applied to value types (structs). Structs cannot inherit from abstract base classes. This allows behavioral contracts/usage guidelines to be applied on value types.

The interviewers are barking up an odd tree. For languages like C# and Java, there is a difference, but in other languages like C++ there is not. OO theory doesn't differentiate the two, merely the syntax of language.

An abstract class is a class with both implementation and interface (pure virtual methods) that will be inherited. Interfaces generally do not have any implementation but only pure virtual functions.

In C# or Java an abstract class without any implementation differs from an interface only in the syntax used to inherit from it and the fact you can only inherit from one.

By implementing interfaces you are achieving composition ("has-a" relationships) instead of inheritance ("is-a" relationships). That is an important principle to remember when it comes to things like design patterns where you need to use interfaces to achieve a composition of behaviors instead of an inheritance.

These answers are all too long.

• Interfaces are for defining behaviors.

• Abstract classes are for defining a thing itself, including its behaviors. That's why we sometimes create an abstract class with some extra properties inheriting an interface.

This also explains why Java only supports single inheritance for classes but puts no restriction on interfaces. Because a concrete object can not be different things, but it can have different behaviors.

For .Net,

Your answer to The second interviewer is also the answer to the first one... Abstract classes can have implementation, AND state, interfaces cannot...

EDIT: On another note, I wouldn't even use the phrase 'subclass' (or the 'inheritance' phrase) to describe classes that are 'defined to implement' an interface. To me, an interface is a definition of a contract that a class must conform to if it has been defined to 'implement' that interface. It does not inherit anything... You have to add everything yourself, explicitly.

tl;dr; When you see “Is A” relationship use inheritance/abstract class. when you see “has a” relationship create member variables. When you see “relies on external provider” implement (not inherit) an interface.

Interview Question: What is the difference between an interface and an abstract class? And how do you decide when to use what? I mostly get one or all of the below answers: Answer 1: You cannot create an object of abstract class and interfaces.

ZK (That’s my initials): You cannot create an object of either. So this is not a difference. This is a similarity between an interface and an abstract class. Counter Question: Why can’t you create an object of abstract class or interface?

Answer 2: Abstract classes can have a function body as partial/default implementation.

ZK: Counter Question: So if I change it to a pure abstract class, marking all the virtual functions as abstract and provide no default implementation for any virtual function. Would that make abstract classes and interfaces the same? And could they be used interchangeably after that?

Answer 3: Interfaces allow multi-inheritance and abstract classes don’t.

ZK: Counter Question: Do you really inherit from an interface? or do you just implement an interface and, inherit from an abstract class? What’s the difference between implementing and inheriting? These counter questions throw candidates off and make most scratch their heads or just pass to the next question. That makes me think people need help with these basic building blocks of Object-Oriented Programming. The answer to the original question and all the counter questions is found in the English language and the UML. You must know at least below to understand these two constructs better.

Common Noun: A common noun is a name given “in common” to things of the same class or kind. For e.g. fruits, animals, city, car etc.

Proper Noun: A proper noun is the name of an object, place or thing. Apple, Cat, New York, Honda Accord etc.

Car is a Common Noun. And Honda Accord is a Proper Noun, and probably a Composit Proper noun, a proper noun made using two nouns.

Coming to the UML Part. You should be familiar with below relationships:

• Is A
• Has A
• Uses

Let’s consider the below two sentences. - HondaAccord Is A Car? - HondaAccord Has A Car?

Which one sounds correct? Plain English and comprehension. HondaAccord and Cars share an “Is A” relationship. Honda accord doesn’t have a car in it. It “is a” car. Honda Accord “has a” music player in it.

When two entities share the “Is A” relationship it’s a better candidate for inheritance. And Has a relationship is a better candidate for creating member variables. With this established our code looks like this:

abstract class Car
{
   string color;
   int speed;
}
class HondaAccord : Car
{
   MusicPlayer musicPlayer;
}

Now Honda doesn't manufacture music players. Or at least it’s not their main business.

So they reach out to other companies and sign a contract. If you receive power here and the output signal on these two wires it’ll play just fine on these speakers.

This makes Music Player a perfect candidate for an interface. You don’t care who provides support for it as long as the connections work just fine.

You can replace the MusicPlayer of LG with Sony or the other way. And it won’t change a thing in Honda Accord.

Why can’t you create an object of abstract classes?

Because you can’t walk into a showroom and say give me a car. You’ll have to provide a proper noun. What car? Probably a honda accord. And that’s when a sales agent could get you something.

Why can’t you create an object of an interface? Because you can’t walk into a showroom and say give me a contract of music player. It won’t help. Interfaces sit between consumers and providers just to facilitate an agreement. What will you do with a copy of the agreement? It won’t play music.

Why do interfaces allow multiple inheritance?

Interfaces are not inherited. Interfaces are implemented. The interface is a candidate for interaction with the external world. Honda Accord has an interface for refueling. It has interfaces for inflating tires. And the same hose that is used to inflate a football. So the new code will look like below:

abstract class Car
{
    string color;
    int speed;
}
class HondaAccord : Car, IInflateAir, IRefueling
{
    MusicPlayer musicPlayer;
}

And the English will read like this “Honda Accord is a Car that supports inflating tire and refueling”.

Interfaces are light weight way to enforce a particular behavior. That is one way to think of.

1) An interface can be seen as a pure Abstract Class, is the same, but despite this, is not the same to implement an interface and inheriting from an abstract class. When you inherit from this pure abstract class you are defining a hierarchy -> inheritance, if you implement the interface you are not, and you can implement as many interfaces as you want, but you can only inherit from one class.

2) You can define a property in an interface, so the class that implements that interface must have that property.

For example:

  public interface IVariable
  {
      string name {get; set;}
  }

The class that implements that interface must have a property like that.

From another answer of mine, mostly dealing with when to use one versus the other:

In my experience, interfaces are best used when you have several classes which each need to respond to the same method or methods so that they can be used interchangeably by other code which will be written against those classes' common interface. The best use of an interface is when the protocol is important but the underlying logic may be different for each class. If you would otherwise be duplicating logic, consider abstract classes or standard class inheritance instead.

Abstract class deals with efficiently packaging the class functionality whereas interface is for intention/contract/communication and is supposed to be shared with other classes/modules.

Using abstract classes as both contract and (partial) contract implementer violates SRP. Using abstract class as a contract (dependency) puts restriction on creating multiple abstract classes for better re-usability.

In the sample below, using abstract class as a contract to OrderManager would create issues as we have two different ways of processing the orders - based on customer type and category (Customer can be either direct or indirect or gold or silver). Hence interface is used for contract and abstract class is used for different workflow enforcement

public interface IOrderProcessor
{
    bool Process(string orderNumber);
}

public abstract class CustomerTypeOrderProcessor: IOrderProcessor
{
    public bool Process(string orderNumber) => IsValid(orderNumber) ? ProcessOrder(orderNumber) : false;

    protected abstract bool ProcessOrder(string orderNumber);

    protected abstract bool IsValid(string orderNumber);
}

public class DirectCustomerOrderProcessor : CustomerTypeOrderProcessor
{
    protected override bool IsValid(string orderNumber) => string.IsNullOrEmpty(orderNumber); 

    protected override bool ProcessOrder(string orderNumber) => true; 
}

public class InDirectCustomerOrderProcessor : CustomerTypeOrderProcessor
{
    protected override bool IsValid(string orderNumber) => orderNumber.StartsWith("EX");

    protected override bool ProcessOrder(string orderNumber) => true;
}

public abstract class CustomerCategoryOrderProcessor : IOrderProcessor
{
    public bool Process(string orderNumber) => ProcessOrder(GetDiscountPercentile(orderNumber), orderNumber);

    protected abstract int GetDiscountPercentile(string orderNumber);

    protected abstract bool ProcessOrder(int discount, string orderNumber);
}

public class GoldCustomer : CustomerCategoryOrderProcessor
{
    protected override int GetDiscountPercentile(string orderNumber) => 15;

    protected override bool ProcessOrder(int discount, string orderNumber) => true;

}

public class SilverCustomer : CustomerCategoryOrderProcessor
{
    protected override int GetDiscountPercentile(string orderNumber) => 10;

    protected override bool ProcessOrder(int discount, string orderNumber) => true;

}

public class OrderManager
{
    private readonly IOrderProcessor _orderProcessor;// Not CustomerTypeOrderProcessor or CustomerCategoryOrderProcessor 

    //Using abstract class here would create problem as we have two different abstract classes
    public OrderManager(IOrderProcessor orderProcessor) => _orderProcessor = orderProcessor;
}

Interface:

Does not contain Impleamentations Interface can inherit from a number of interfaces(Multiple inheritance supported) Members are automatically public May contain properties, methods, events and indexers

Abstract Class:

May/mayn't contain Implementations; At least one member will not be implemented. A Class may inherit from a single Base class; multiple inheritance not allowed. Members have access modifiers May contain fields, properties, constructors, destructors, methods, events and indexers

These are the key differences between Abstract Class and Interfaces.