Let us have a look at the eventing or observer pattern. This pattern is based on the raising and handling of events. Events are a mechanism for the communication between object that allow us to build loosely coupled applications that can be easily extended.

At a high level this pattern functions as follows:
An object (known as the Publisher) defines a contract (delegate method) to which other objects (knows as Subscribers) must comply in order to be notified of a certain condition occurring in the Publisher object. This is achieved as follows, when a certain state is reached in the Publisher object an event will be raised, which will then trigger all the Subscriber objects to react by executing the method defined within the individual Subscriber objects that matches the delegate method as defined in the Publisher.

Now let us have a look at an example:

First let us create the Publisher, here we have to do three things:
1. Define a delegate which will act as the contract between the Publisher and Subscribers.
2. Define an event based on the delegate.
3. Raise the defined event.

public class KillerRobotPublisher
{
	public delegate void KillerRobotEventHandler(object source, EventArgs args);

	public event KillerRobotEventHandler KillerRobotAction;

	protected virtual void OnKillerRobotAction()
	 {
		if(KillerRobotAction!=null) //check if there are any subscribers
			{
				KillerRobotAction(this,EventArgs.Empty);
			}
	 }

	public void DoKillerRobotAction(string action)
	 {
		Console.WriteLine(“Killer Robot is doing “ +action);
		//add additional action logic here

		OnKillerRobotAction();
	}
}

Next let us create a Subscriber to subscribe to the KillerRobotAction event on the Publisher.

public class KillerRobotSubscriber
{
	public void OnKillerRobotAction(object source, EventArgs args)
     //this method conforms to the delegate defined in the Publisher
	 {
		Console.WriteLine(“The Killer Robot did something!”);
	 }
}

Now lastly let us create a basic application to instantiate the objects and to subscribe the Subscriber to the Publisher:

class Program
{
	static void Main(string[] args)
	{
		var killerRobotPub = new KillerRobotPublisher(); //Instantiate Publisher Object
		var killerRobotSub = new KillerRobotSubscriber(); //Instantiate Subscriber Object 

		killerRobotPub.KillerRobotAction += killerRobotSub.OnKillerRobotAction;
		//Subscribe the KillerRobotSubscriber to the KillerRobotAction event on the KillerRobotPublisher.

		killerRobotPub.DoKillerRobotAction(“A Dance”);
	}
}

The console output of this application will look like this:

Extension methods allow us to add methods to existing classes without changing the class’ source code, nor by inheriting from the class. Extension methods are more relevant when wanting to add a method to a class from which one cannot inherit, such as sealed classes.
Just note, you cannot use extension methods with a static class as extension methods require an instance variable for an object in order to be utilised.
Let us look at an example, here we will add an extension method to the string class, which is a sealed class (i.e. you cannot inherit from this class).

class Program
{
	static void Main(string[] args)
	{
		string sentence=  “This is something a person would say.”;

		var robotSentence = sentence.ToRobot();
		Console.WriteLine(robotSentence);
	}
}

public static class StringExtensions
{
	public static string ToRobot(this string str)
	{
		if(String.IsNullOrEmpty(str)) return str;
		var words = str.Split(‘ ‘);
		var robotStr = String.Empty;

		foreach(var word in words)
		{
			if(word.Length > 4)
			{
				robotStr+=“BEEP “;
			} else {
				robotStr+=“BOOP ”;
			}
		}

		return robotStr.Trim();
	}
}

Also note that extension methods must be non-generic static methods defined in a static class.

A delegate is a form of a type-safe function pointer. More simply put, it is an object that knows how to call a method, i.e. a reference to a method. Delegates are useful as they assist us in writing flexible and extendable applications.

Delegates are useful when using the eventing design pattern also known as the observer pattern. The eventing or observer pattern consists of an object, called the subject or publisher, which maintains a list of dependent objects, called observers or subscribers, and notifies them automatically of a state change in the subject-object by raising an event and then calling a method on the observer objects by using a delegate. We will cover this pattern in more detail when I cover events in a separate post.

For now, let us simply focus on delegates.

So, let us look at an example:

public class KillerRobot
{
	public string Name { get; set; }
}

public class KillerRobotActuator
{
	public delegate void KillerRobotActionHandler(KillerRobot robot); 	//Declare delegate which will act as a
															//signature for methods that can be
															//referenced.

	public void DoAction(string robotName, KillerRobotActionHandler actionHandler)
	{
		var robot = new KillerRobot {Name = robotName	};
		actionHandler(robot);
	}
}

public class RobotActions
{
	public void RobotTalk(KillerRobot robot)
	{
		Console.WriteLine(“{0} is Talking.”, robot.Name);
	}

	public void RobotDance(KillerRobot robot)
	{
		Console.WriteLine(“{0} is Dancing.”, robot.Name);
	}
}

class Program
{
	static void Main(string[] args)
	{
		var robotActuator = new KillerRobotActuator();
		var robotActions = new RobotActions();
		KillerRobotActuator.KillerRobotActionHandler actionHandler = robotActions.RobotTalk;
		actionHandler += robotActions.RobotDance;
		actionHandler += RobotPowerOff;

		robotActuator.DoAction(“The Geek”,actionHandler);
	}

	static void RobotPowerOff(KillerRobot robot)
	{
		Console.WriteLine(“{0} has turned off.”, robot.Name);
	}
}

So, in this basic example we have a class KillerRobotActuator which declares a delegate:

public delegate void KillerRobotActionHandler(KillerRobot robot);

Any method that complies to this signature can then be added to this delegate and in the DoAction method where the following is executed:

actionHandler(robot);

All the methods that have been added to the actionHandler will then be executed.
We can see in the Main method of the Program class that a new KillerRobotActionHandler is declared and three method references are added to it as below:

KillerRobotActuator.KillerRobotActionHandler actionHandler = robotActions.RobotTalk;
actionHandler += robotActions.RobotDance;
actionHandler += RobotPowerOff;

And then finally the DoAction method on KillerRobotActuator is executed, passing in the above declared actionHandler containing references to the three methods:

robotActuator.DoAction(“The Geek”,actionHandler);

All three the methods that are referenced by the actionHandler comply to the signature defined in the delegate declaration in the KillerRobotActuator class. i.e. a void return type and an input parameter of type KillerRobot.
It is also worth mentioning that of the methods referenced, two are contained in a separate class RobotActions and the third is a static method declared in the Program class, so methods from multiple different class locations can be added as long as they comply to the signature of the delegate declares.

Generics refer to the method of creating classes and methods in a way that defers the specification of the type or types associated with the class or method until it is declared and instantiated.

What this means in plain english is that you can define a single class or method that can be utilised with multiple types, thus resulting in less and tidier code.

So let us have a look at an example. Firstly let us look at some code that is not generic. Here we have a class which consist of a list of integers, it has a default constructor as well as two methods, one to add an integer to the list and the other to return the sum of all the integers in the list.

public class KillerRoboticsIntList
{
   public List<int> IntList { get; set; }

    public KillerRoboticsIntList()
    {
        IntList = new List<int>();
    }

    public void Add(int item)
    {
        IntList.Add(item);
    }

    public int Sum()
   {
	int sum;
	foreach(int i in IntList)
	{
	 sum += i;
	}
	return sum;
   }

}

This is a very basic example and can be utilised as follows:

class Program
        {
            public static void Main()
            {
                KillerRoboticsIntList iList = new KillerRoboticsIntList();
		        iList.Add(1);
		        iList.Add(2);
		        iList.Add(3);
		        iList.Add(4);
		        int sum = iList.Sum();
                Console.WriteLine(“Sum = {0}.", sum);

            }
        }

This is very straight forward, however if I would like to use this class with a data type other than int (for example a float or a double) I would need to define a new class for each data type. This can be overcome by utilising Generics. Below is an example of how this can be implemented:

public class KillerRoboticsGenericList<T>
{
   public List GenList<T> { get; set; }

    public KillerRoboticsGenericList()
    {
        GenList = new List<T>();
    }

    public void Add(T item)
    {
        GenList.Add(item);
    }

    public T Sum()
   {
	T sum;
	foreach(T i in GenList)
	{
	 sum += i;
	}
	return sum;
   }

}

The above class can now be defined for various types as seen below:

class Program
        {
            public static void Main()
            {
              KillerRoboticsIntList<int> intList = new KillerRoboticsIntList<int>();
		        intList.Add(1);
		        intList.Add(2);
		        intList.Add(3);
		        intList.Add(4);
		        int sum = intList.Sum();
                Console.WriteLine(“Int Sum = {0}.", sum);

		        KillerRoboticsIntList<double> = new KillerRoboticsIntList<double>();
		        doubleList.Add(0.1);
		        doubleList.Add(2.2);
		        doubleList.Add(3.0);
		        doubleList.Add(1.4);
		        double dSum = doubleList.Sum();
                Console.WriteLine(“Double Sum = {0}.", dSum);

            }
        }

In some cases you might want to put a restriction on the types that can be utilised to implement a generic class of method, this can be achieved by utilising constraints. For example the class we defined above can be restricted to types that implement the IComparable interface by simple changing the first line as follows:

public class KillerRoboticsGenericList<T> where T : IComparable

Some additional examples of constraints are:

where T : Product

This restricts the type of T to an implementation of the class Product or any of its child classes.

where T : struct

This restricts the type of T to the value type struct.

where T : new()

This restricts the type of T to an object with a default constructor.

Multiple Constraints can also be appled at the same time, for example:

 
public class KillerRoboticsGenericList<T> where T : IComparable, new() 

A Generic class can be defined to utilise multiple types, for example:

 
public class KillerRoboticsGenericList<T,U,V>  

It is also worth mentioning that the default c# List class used in the above examples also utilises generics, along with all the other predefined c# collections contained in System.Collections.Generic.

I hope this post has been useful and I will be posting on some additional C# topics, such as Delegates, Lamda Expressions, LINQ, Extension Methods, etc. over the next few months.