We Keep Coding

Listen for event and invoke callback, based on specification?

I am currently building out a custom task manager and I'm wondering if it's possible to tell the task manager to listen for a specific event (OnSomething below), and then invoke a callback method when the task raises that event. However, mentally I can't see how it's possible to listen for an event that doesn't exist at the base class level. For example, I have a base class that contains basic information about the task called CustomTask:
public abstract class CustomTask {
public bool IsRunning { get; private set; } = false;
public void Start() {
IsRunning = true;
DoSomething();
IsRunning = false;
}
protected abstract void DoSomething();
}
For the sake of SO readers, I've simplified the definition, but you get the gist of it. It contains basic details, a few methods for starting and canceling, provides basic state management (simplified IsRunning here), etc.
I then have custom tasks that derive from CustomTask, in this case, let's focus on a sample task called CustomTaskA. It contains a definition for an event called OnSomething, which someone, somewhere may want to listen for:
public sealed class CustomTaskA : CustomTask {
protected override void DoSomething() => RaiseOnSomething(this, new EventArgs());
public event EventHandler<EventArgs> OnSomething;
private void RaiseOnSomething(object sender, EventArgs e) => OnSomething?.Invoke(sender, e);
}
Now, the CustomTaskManager registers tasks, tracks them via Guid, manages them and more, but for simplicity:
public sealed class CustomTaskManager {
// Singleton setup.
private static CustomTaskManager _instance = new CustomTaskManager();
public static CustomTaskManager Instance {
get {
// Simplified for SO.
if (_instance == null)
_instance = new CustomTaskManager();
return;
}
}
// Collection of tasks.
private Dictionary<Guid, CustomTask> _tasks = new Dictionary<Guid, CustomTask>();
// Register and start a task.
public bool TryRegisterAndStartTask(CustomTask task, out Guid taskId) {
taskId = Guid.Empty;
try {
// Register task.
taskId = Guid.NewGuid();
_tasks.Add(taskId, task);
// Listen for events.
// Start task.
task.Start();
} catch (Exception e) {
// Log exception.
}
return false;
}
}
When registering and starting a task, I'd like to tell the task manager I want to listen for OnSomething, and if OnSomething is invoked, I want the task manager to call a method OnSomethingWasRaised. For example:
TaskManager.Instance.TryRegisterAndStartTask(task, out Guid taskId, task.OnSomething, OnSomethingWasRaised);
private static void OnSomethingWasRaised(object sender, EventArgs e) {
Console.WriteLine("Woohoo!");
}
I know the specifying and invoking a callback method is entirely possible, and listening for events is plausible with reflection.
Is there a way (with or without using reflection) to listen for a specified event defined on a derived object and then invoke a specified callback method?
NOTE: Please excuse any syntactical errors as I hand-typed the snippets to keep them minimal.

Problem with (proposed) approach like this:
TryRegisterAndStartTask(task, out Guid taskId, task.OnSomething, OnSomethingWasRaised);
is that you cannot pass event as argument, or store it in variable, because event is just a set of two methods (add and remove), just like property is a set of two methods get and set.
You can of course change event to "raw" delegate:
public EventHandler<EventArgs> OnSomething;
This one you can pass by reference:
public bool TryRegisterAndStartTask(CustomTask task, ref EventHandler<EventArgs> del, EventHandler<EventArgs> sub, out Guid taskId) {
taskId = Guid.Empty;
// subscribe
del += sub;
...
}
CustomTaskManager.Instance.TryRegisterAndStartTask(task, ref task.OnSomething, OnSomethingWasRaised, out var taskId);
But that's usually not a good idea, since you are losing private scope of events - with events one can only add\remove delegates, with raw delegate anyone can do anything, like invoking or setting to null.
If regular event stays - that means reflection is the only way to achieve your goal, and even worse - you'll have to reference to the event you want to subscribe to by string name, not by an actual reference, though you can use nameof(task.OnSomething). Then, you are losing compile time validation of subscription delegate type. Say you want to subscribe to event Action Something but passing Func<string> delegate there. It will compile fine with reflection approach, and fail only at runtime.
Still if you insist that will look something like this:
public bool TryRegisterAndStartTask(CustomTask task, string eventName, Delegate sub, out Guid taskId) {
taskId = Guid.Empty;
// subscribe
var ev = task.GetType().GetEvent(eventName, BindingFlags.Public | BindingFlags.Instance);
var addMethod = ev.GetAddMethod(); // this can be null or private by the way
addMethod.Invoke(task, new [] {sub});
...
}
And called like this:
var task = new CustomTaskA();
EventHandler<EventArgs> handler = OnSomethingWasRaised;
CustomTaskManager.Instance.TryRegisterAndStartTask(task, nameof(task.OnSomething), handler, out var taskId);
Ugly, unsafe, and not worth it in your scenario, in my opinion.

Monotouch Unsubscribe From Events

I have been using Events instead of delegates in my MonoTouch project, typically using this pattern (iPhone app using Storyboards):
I call PerformSegue to present a new View Controller, and in PrepareForSegue method, I set the View Controller's properties and subscribe to its events using a lambda expression as such:
public override void PrepareForSegue (UIStoryboardSegue segue, NSObject sender)
{
if (segue.Identifier.Equals("NextViewControllerSegue")) {
using (MyNextViewController destinationVC = segue.DestinationViewController as MyNextViewController) {
destinationVC.SomeProperty = "some value";
destinationVC.Cancelled += (s, e) => {
this.DismissViewController(false, null);
};
}
}
}
(as an aside, I chain these Cancelled events where I need to close a hierarchy of View Controllers - whether this is good or bad is a question for another day)
My question is: I recently discovered that if you don't unsubscribe from an object's events, that object is not Garbage Collected. So I have changed the above code to:
public override void PrepareForSegue (UIStoryboardSegue segue, NSObject sender)
{
if (segue.Identifier.Equals("NextViewControllerSegue")) {
using (MyNextViewController destinationVC = segue.DestinationViewController as MyNextViewController) {
destinationVC.SomeProperty = "some value";
destinationVC.Cancelled += Cancel;
}
}
}
protected void Cancel (object sender, EventArgs e)
{
(sender as MyNextViewController).Cancelled -= Cancel;
this.DismissViewController(false, null);
}
My question is this: is this pattern a good way to go about things? And will approach 2 (unsubscribing from the event in the event delegate) work? I am not sure where else to unsubscribe. Or should I move everything to use a notification pattern instead (as suggested here: Should I prefer NSNotificactionCenter or .NET events when using Monotouch?)?

In your first example, the event will cause the destinationVC to have a reference to this. Since this is most likely lower in the stack than destinationVC in a navigation controller, you don't really have an issue because destinationVC doesn't live as long as this.
Your second example isn't needed in your example, but it would be if:
A controller subscribes to an event on an object that will be around for a long time, like a Model/ViewModel class
A controller subscribes to an event on a previous controller in a navigation controller stack (same goes for modal controllers)

.net default event handler

In my product I need process wide events. For that I used code like this:
public class Global
{
public static event EventHandler<MyEventArgs> Message;
public static void ShowMessage();
}
Now let's say I have a WinForms user interface. In form's code I will subscribe to this event and handle it in some default way (eg. by using System.Windows.Forms.MessageBox.Show() method). Now the question is how do I allow user to create derived form and override my default Message event handler implementation?
Just subscribing to the event for the second time with custom implementation doesn't solve the problem (both event handlers would be executed and potentially two message boxes shown). The options I see are either:
//call OnSubscribeToMessageEvent() from either form's constructor or OnLoad event handler
protected virtual void OnSubscribeToMessageEvent()
{
Global.Message += new EventHandler<MyEventArgs>(Global_Message);
}
private void Global_Message(object sender, MyEventArgs e)
{
//my default implementation
}
or
//subscribe in either form's constructor or OnLoad event handler
protected virtual void Global_Message(object sender, MyEventArgs e)
{
//my default implementation
}
Which version is better and why? Or maybe there are any other options?

I still have some doubts as I have never seen such a design pattern in any .NET library
Yes, you're right to worry about this. These kind of event subscriptions are very fickle, the event source always outlives the subscriber. There's only one class in the framework I know that does this, SystemEvents. The problem is that every subscriber has to very carefully unsubscribe itself when its lifetime ends or the object will stay referenced forever. A memory leak that's very hard to diagnose.
A better pattern here is to use an interface. Let's declare one:
public class MyEventArgs { /* etc.. */ }
public interface IGlobalNotification {
event EventHandler Disposed;
void OnMessage(MyEventArgs arg);
}
Now you can have a form implement the interface:
public partial class Form1 : Form, IGlobalNotification {
public Form1() {
InitializeComponent();
GlobalMessages.Register(this);
}
void IGlobalNotification.OnMessage(MyEventArgs arg) {
// do something
}
}
The Register method registers the form with the GlobalMessages class, the Dispose event ensures that the class can detect that the form is dying:
public static class GlobalMessages {
public static void Register(IGlobalNotification listener) {
listener.Disposed += delegate { listeners.Remove(listener); };
listeners.Add(listener);
}
public static void Notify(MyEventArgs arg) {
foreach (var listener in listeners) listener.OnMessage(arg);
}
private static List<IGlobalNotification> listeners = new List<IGlobalNotification>();
}
Call GlobalMessages.Notify() to get the OnMessage() method to run in all live form instances. The major advantage of this approach is that a client programmer can never screw up.

I would let the derived class override the Global_Message. The subscription to the event is generic and why would you want to implement it in every child again? It also gives you the option to call base.Global_Message(sender, e) in case your child class just wants to add some decoration to it and use the default behaviour otherwise.

I would prefer your second example, as that way, classes that extend your base class only have to override one method and do not have to remove the handler added by the base class from the event.

The key is adding the virtual keyword, so that a derived type can overide the method and the method they created will be called instead.
//subscribe in either form's constructor or OnLoad event handler
protected virtual void Global_Message(object sender, MyEventArgs e)
{
//my default implementation
}

Now that you've added virtual to both, I'd go with the first and override the one that subscribes to the event, if they didn't want the event subscribed to.
Though there is another option, call it #3.
protected EventHandler GlobalMessageEvent = new EventHandler<MyEventArgs>(Global_Message);
protected virtual void OnSubscribeToMessageEvent()
{
// this could be done in the Form_Load() or constructor instead.
Global.Message += GlobalMessageEvent;
}
Then potentially an inherited class could do somewhere: (note the -=)
{
Global.Message -= GlobalMessageEvent;
}

Are there pitfalls to using static class/event as an application message bus

I have a static generic class that helps me move events around with very little overhead:
public static class MessageBus<T> where T : EventArgs
{
public static event EventHandler<T> MessageReceived;
public static void SendMessage(object sender, T message)
{
if (MessageReceived != null)
MessageReceived(sender, message);
}
}
To create a system-wide message bus, I simply need to define an EventArgs class to pass around any arbitrary bits of information:
class MyEventArgs : EventArgs
{
public string Message { get; set; }
}
Anywhere I'm interested in this event, I just wire up a handler:
MessageBus<MyEventArgs>.MessageReceived += (s,e) => DoSomething();
Likewise, triggering the event is just as easy:
MessageBus<MyEventArgs>.SendMessage(this, new MyEventArgs() {Message="hi mom"});
Using MessageBus and a custom EventArgs class lets me have an application wide message sink for a specific type of message. This comes in handy when you have several forms that, for example, display customer information and maybe a couple forms that update that information. None of the forms know about each other and none of them need to be wired to a static "super class".
I have a couple questions:
fxCop complains about using static methods with generics, but this is exactly what I'm after here. I want there to be exactly one MessageBus for each type of message handled. Using a static with a generic saves me from writing all the code that would maintain the list of MessageBus objects.
Are the listening objects being kept "alive" via the MessageReceived event?
For instance, perhaps I have this code in a Form.Load event:
MessageBus<CustomerChangedEventArgs>.MessageReceived += (s,e) => DoReload();
When the Form is Closed, is the Form being retained in memory because MessageReceived has a reference to its DoReload method? Should I be removing the reference when the form closes:
MessageBus<CustomerChangedEventArgs>.MessageReceived -= (s,e) => DoReload();

Well, yes, you should, but if you use the lambda syntax as you've done in your example I think it won't work (by which I mean, the handler will not be de-registered successfully).
Someone correct me if I'm wrong, but I believe this is true because using the lambda syntax effectively creates a new EventHandler<CustomerChangedEventArgs> object, with its own place in memory. When you try to remove this handler, using the lambda syntax again, this creates yet another new EventHandler<CustomerChangedEventArgs> object, which is not equal to the first one you created; and so the first one never gets de-registered.
Sadly, I think you'll need to actually define a method like this:
DoReload(object sender, CustomerChangedEventArgs e) {
DoReload(); // your original overload, which doesn't actually care
// about the sender and e parameters
}
This way you can do:
MessageBus<CustomerChangedEventArgs>.MessageReceived += DoReload;
And later:
MessageBus<CustomerChangedEventArgs>.MessageReceived -= DoReload;

Yes, there are problems. Your event handlers will cause the form object to stay referenced, you have to explicitly unregister the event handlers. The lambdas make this impossible, you'll have to write an explicit handler.
This pattern has a name, "Event Broker service". It is part of the Composite UI Application Block, published by Microsoft's Pattern and Practices team. Beg, borrow and steal (if not use) what you can from this.

You could use weak references to store the event handlers. That way, unhooked handlers won't prevent garbage collection of the objects.
public static class MessageBus<T> where T : EventArgs
{
private static List<WeakReference> _handlers = new List<WeakReference>();
public static event EventHandler<T> MessageReceived
{
add
{
_handlers.Add(new WeakReference(value));
}
remove
{
// also remove "dead" (garbage collected) handlers
_handlers.RemoveAll(wh => !wh.IsAlive || wh.Target.Equals(value));
}
}
public static void SendMessage(object sender, T message)
{
foreach(var weakHandler in _handlers)
{
if (weakHandler.IsAlive)
{
var handler = weakHandler.Target as EventHandler<T>;
handler(sender, message);
}
}
}
}

Event vs Delegates [duplicate]

What are the differences between delegates and an events? Don't both hold references to functions that can be executed?

An Event declaration adds a layer of abstraction and protection on the delegate instance. This protection prevents clients of the delegate from resetting the delegate and its invocation list and only allows adding or removing targets from the invocation list.

To understand the differences you can look at this 2 examples
Example with Delegates (in this case, an Action - that is a kind of delegate that doesn't return a value)
public class Animal
{
public Action Run {get; set;}
public void RaiseEvent()
{
if (Run != null)
{
Run();
}
}
}
To use the delegate, you should do something like this:
Animal animal= new Animal();
animal.Run += () => Console.WriteLine("I'm running");
animal.Run += () => Console.WriteLine("I'm still running") ;
animal.RaiseEvent();
This code works well but you could have some weak spots.
For example, if I write this:
animal.Run += () => Console.WriteLine("I'm running");
animal.Run += () => Console.WriteLine("I'm still running");
animal.Run = () => Console.WriteLine("I'm sleeping") ;
with the last line of code, I have overridden the previous behaviors just with one missing + (I have used = instead of +=)
Another weak spot is that every class which uses your Animal class can invoke the delegate directly. For example, animal.Run() or animal.Run.Invoke() are valid outside the Animal class.
To avoid these weak spots you can use events in c#.
Your Animal class will change in this way:
public class ArgsSpecial : EventArgs
{
public ArgsSpecial (string val)
{
Operation=val;
}
public string Operation {get; set;}
}
public class Animal
{
// Empty delegate. In this way you are sure that value is always != null
// because no one outside of the class can change it.
public event EventHandler<ArgsSpecial> Run = delegate{}
public void RaiseEvent()
{
Run(this, new ArgsSpecial("Run faster"));
}
}
to call events
Animal animal= new Animal();
animal.Run += (sender, e) => Console.WriteLine("I'm running. My value is {0}", e.Operation);
animal.RaiseEvent();
Differences:
You aren't using a public property but a public field (using events, the compiler protects your fields from unwanted access)
Events can't be assigned directly. In this case, it won't give rise to the previous error that I have showed with overriding the behavior.
No one outside of your class can raise or invoke the event. For example, animal.Run() or animal.Run.Invoke() are invalid outside the Animal class and will produce compiler errors.
Events can be included in an interface declaration, whereas a field cannot
Notes:
EventHandler is declared as the following delegate:
public delegate void EventHandler (object sender, EventArgs e)
it takes a sender (of Object type) and event arguments. The sender is null if it comes from static methods.
This example, which uses EventHandler<ArgsSpecial>, can also be written using EventHandler instead.
Refer here for documentation about EventHandler

In addition to the syntactic and operational properties, there's also a semantical difference.
Delegates are, conceptually, function templates; that is, they express a contract a function must adhere to in order to be considered of the "type" of the delegate.
Events represent ... well, events. They are intended to alert someone when something happens and yes, they adhere to a delegate definition but they're not the same thing.
Even if they were exactly the same thing (syntactically and in the IL code) there will still remain the semantical difference. In general I prefer to have two different names for two different concepts, even if they are implemented in the same way (which doesn't mean I like to have the same code twice).

Here is another good link to refer to.
http://csharpindepth.com/Articles/Chapter2/Events.aspx
Briefly, the take away from the article - Events are encapsulation over delegates.
Quote from article:
Suppose events didn't exist as a concept in C#/.NET. How would another class subscribe to an event? Three options:
A public delegate variable
A delegate variable backed by a property
A delegate variable with AddXXXHandler and RemoveXXXHandler methods
Option 1 is clearly horrible, for all the normal reasons we abhor public variables.
Option 2 is slightly better, but allows subscribers to effectively override each other - it would be all too easy to write someInstance.MyEvent = eventHandler; which would replace any existing event handlers rather than adding a new one. In addition, you still need to write the properties.
Option 3 is basically what events give you, but with a guaranteed convention (generated by the compiler and backed by extra flags in the IL) and a "free" implementation if you're happy with the semantics that field-like events give you. Subscribing to and unsubscribing from events is encapsulated without allowing arbitrary access to the list of event handlers, and languages can make things simpler by providing syntax for both declaration and subscription.

What a great misunderstanding between events and delegates!!! A delegate specifies a TYPE (such as a class, or an interface does), whereas an event is just a kind of MEMBER (such as fields, properties, etc). And, just like any other kind of member an event also has a type. Yet, in the case of an event, the type of the event must be specified by a delegate. For instance, you CANNOT declare an event of a type defined by an interface.
Concluding, we can make the following Observation: the type of an event MUST be defined by a delegate. This is the main relation between an event and a delegate and is described in the section II.18 Defining events of ECMA-335 (CLI) Partitions I to VI:
In typical usage, the TypeSpec (if present) identifies a delegate whose signature matches the arguments passed to the event’s fire method.
However, this fact does NOT imply that an event uses a backing delegate field. In truth, an event may use a backing field of any different data structure type of your choice. If you implement an event explicitly in C#, you are free to choose the way you store the event handlers (note that event handlers are instances of the type of the event, which in turn is mandatorily a delegate type---from the previous Observation). But, you can store those event handlers (which are delegate instances) in a data structure such as a List or a Dictionary or any other else, or even in a backing delegate field. But don’t forget that it is NOT mandatory that you use a delegate field.

NOTE: If you have access to C# 5.0 Unleashed, read the "Limitations on Plain Use of Delegates" in Chapter 18 titled "Events" to understand better the differences between the two.
It always helps me to have a simple, concrete example. So here's one for the community. First I show how you can use delegates alone to do what Events do for us. Then I show how the same solution would work with an instance of EventHandler. And then I explain why we DON'T want to do what I explain in the first example. This post was inspired by an article by John Skeet.
Example 1: Using public delegate
Suppose I have a WinForms app with a single drop-down box. The drop-down is bound to an List<Person>. Where Person has properties of Id, Name, NickName, HairColor. On the main form is a custom user control that shows the properties of that person. When someone selects a person in the drop-down the labels in the user control update to show the properties of the person selected.
Here is how that works. We have three files that help us put this together:
Mediator.cs -- static class holds the delegates
Form1.cs -- main form
DetailView.cs -- user control shows all details
Here is the relevant code for each of the classes:
class Mediator
{
public delegate void PersonChangedDelegate(Person p); //delegate type definition
public static PersonChangedDelegate PersonChangedDel; //delegate instance. Detail view will "subscribe" to this.
public static void OnPersonChanged(Person p) //Form1 will call this when the drop-down changes.
{
if (PersonChangedDel != null)
{
PersonChangedDel(p);
}
}
}
Here is our user control:
public partial class DetailView : UserControl
{
public DetailView()
{
InitializeComponent();
Mediator.PersonChangedDel += DetailView_PersonChanged;
}
void DetailView_PersonChanged(Person p)
{
BindData(p);
}
public void BindData(Person p)
{
lblPersonHairColor.Text = p.HairColor;
lblPersonId.Text = p.IdPerson.ToString();
lblPersonName.Text = p.Name;
lblPersonNickName.Text = p.NickName;
}
}
Finally we have the following code in our Form1.cs. Here we are Calling OnPersonChanged, which calls any code subscribed to the delegate.
private void comboBox1_SelectedIndexChanged(object sender, EventArgs e)
{
Mediator.OnPersonChanged((Person)comboBox1.SelectedItem); //Call the mediator's OnPersonChanged method. This will in turn call all the methods assigned (i.e. subscribed to) to the delegate -- in this case `DetailView_PersonChanged`.
}
Ok. So that's how you would get this working without using events and just using delegates. We just put a public delegate into a class -- you can make it static or a singleton, or whatever. Great.
BUT, BUT, BUT, we do not want to do what I just described above. Because public fields are bad for many, many reason. So what are our options? As John Skeet describes, here are our options:
A public delegate variable (this is what we just did above. don't do this. i just told you above why it's bad)
Put the delegate into a property with a get/set (problem here is that subscribers could override each other -- so we could subscribe a bunch of methods to the delegate and then we could accidentally say PersonChangedDel = null, wiping out all of the other subscriptions. The other problem that remains here is that since the users have access to the delegate, they can invoke the targets in the invocation list -- we don't want external users having access to when to raise our events.
A delegate variable with AddXXXHandler and RemoveXXXHandler methods
This third option is essentially what an event gives us. When we declare an EventHandler, it gives us access to a delegate -- not publicly, not as a property, but as this thing we call an event that has just add/remove accessors.
Let's see what the same program looks like, but now using an Event instead of the public delegate (I've also changed our Mediator to a singleton):
Example 2: With EventHandler instead of a public delegate
Mediator:
class Mediator
{
private static readonly Mediator _Instance = new Mediator();
private Mediator() { }
public static Mediator GetInstance()
{
return _Instance;
}
public event EventHandler<PersonChangedEventArgs> PersonChanged; //this is just a property we expose to add items to the delegate.
public void OnPersonChanged(object sender, Person p)
{
var personChangedDelegate = PersonChanged as EventHandler<PersonChangedEventArgs>;
if (personChangedDelegate != null)
{
personChangedDelegate(sender, new PersonChangedEventArgs() { Person = p });
}
}
}
Notice that if you F12 on the EventHandler, it will show you the definition is just a generic-ified delegate with the extra "sender" object:
public delegate void EventHandler<TEventArgs>(object sender, TEventArgs e);
The User Control:
public partial class DetailView : UserControl
{
public DetailView()
{
InitializeComponent();
Mediator.GetInstance().PersonChanged += DetailView_PersonChanged;
}
void DetailView_PersonChanged(object sender, PersonChangedEventArgs e)
{
BindData(e.Person);
}
public void BindData(Person p)
{
lblPersonHairColor.Text = p.HairColor;
lblPersonId.Text = p.IdPerson.ToString();
lblPersonName.Text = p.Name;
lblPersonNickName.Text = p.NickName;
}
}
Finally, here's the Form1.cs code:
private void comboBox1_SelectedIndexChanged(object sender, EventArgs e)
{
Mediator.GetInstance().OnPersonChanged(this, (Person)comboBox1.SelectedItem);
}
Because the EventHandler wants and EventArgs as a parameter, I created this class with just a single property in it:
class PersonChangedEventArgs
{
public Person Person { get; set; }
}
Hopefully that shows you a bit about why we have events and how they are different -- but functionally the same -- as delegates.

You can also use events in interface declarations, not so for delegates.

Delegate is a type-safe function pointer. Event is an implementation of publisher-subscriber design pattern using delegate.

An event in .net is a designated combination of an Add method and a Remove method, both of which expect some particular type of delegate. Both C# and vb.net can auto-generate code for the add and remove methods which will define a delegate to hold the event subscriptions, and add/remove the passed in delegagte to/from that subscription delegate. VB.net will also auto-generate code (with the RaiseEvent statement) to invoke the subscription list if and only if it is non-empty; for some reason, C# doesn't generate the latter.
Note that while it is common to manage event subscriptions using a multicast delegate, that is not the only means of doing so. From a public perspective, a would-be event subscriber needs to know how to let an object know it wants to receive events, but it does not need to know what mechanism the publisher will use to raise the events. Note also that while whoever defined the event data structure in .net apparently thought there should be a public means of raising them, neither C# nor vb.net makes use of that feature.

To define about event in simple way:
Event is a REFERENCE to a delegate with two restrictions
Cannot be invoked directly
Cannot be assigned values directly (e.g eventObj = delegateMethod)
Above two are the weak points for delegates and it is addressed in event. Complete code sample to show the difference in fiddler is here https://dotnetfiddle.net/5iR3fB .
Toggle the comment between Event and Delegate and client code that invokes/assign values to delegate to understand the difference
Here is the inline code.
/*
This is working program in Visual Studio. It is not running in fiddler because of infinite loop in code.
This code demonstrates the difference between event and delegate
Event is an delegate reference with two restrictions for increased protection
1. Cannot be invoked directly
2. Cannot assign value to delegate reference directly
Toggle between Event vs Delegate in the code by commenting/un commenting the relevant lines
*/
public class RoomTemperatureController
{
private int _roomTemperature = 25;//Default/Starting room Temperature
private bool _isAirConditionTurnedOn = false;//Default AC is Off
private bool _isHeatTurnedOn = false;//Default Heat is Off
private bool _tempSimulator = false;
public delegate void OnRoomTemperatureChange(int roomTemperature); //OnRoomTemperatureChange is a type of Delegate (Check next line for proof)
// public OnRoomTemperatureChange WhenRoomTemperatureChange;// { get; set; }//Exposing the delegate to outside world, cannot directly expose the delegate (line above),
public event OnRoomTemperatureChange WhenRoomTemperatureChange;// { get; set; }//Exposing the delegate to outside world, cannot directly expose the delegate (line above),
public RoomTemperatureController()
{
WhenRoomTemperatureChange += InternalRoomTemperatuerHandler;
}
private void InternalRoomTemperatuerHandler(int roomTemp)
{
System.Console.WriteLine("Internal Room Temperature Handler - Mandatory to handle/ Should not be removed by external consumer of ths class: Note, if it is delegate this can be removed, if event cannot be removed");
}
//User cannot directly asign values to delegate (e.g. roomTempControllerObj.OnRoomTemperatureChange = delegateMethod (System will throw error)
public bool TurnRoomTeperatureSimulator
{
set
{
_tempSimulator = value;
if (value)
{
SimulateRoomTemperature(); //Turn on Simulator
}
}
get { return _tempSimulator; }
}
public void TurnAirCondition(bool val)
{
_isAirConditionTurnedOn = val;
_isHeatTurnedOn = !val;//Binary switch If Heat is ON - AC will turned off automatically (binary)
System.Console.WriteLine("Aircondition :" + _isAirConditionTurnedOn);
System.Console.WriteLine("Heat :" + _isHeatTurnedOn);
}
public void TurnHeat(bool val)
{
_isHeatTurnedOn = val;
_isAirConditionTurnedOn = !val;//Binary switch If Heat is ON - AC will turned off automatically (binary)
System.Console.WriteLine("Aircondition :" + _isAirConditionTurnedOn);
System.Console.WriteLine("Heat :" + _isHeatTurnedOn);
}
public async void SimulateRoomTemperature()
{
while (_tempSimulator)
{
if (_isAirConditionTurnedOn)
_roomTemperature--;//Decrease Room Temperature if AC is turned On
if (_isHeatTurnedOn)
_roomTemperature++;//Decrease Room Temperature if AC is turned On
System.Console.WriteLine("Temperature :" + _roomTemperature);
if (WhenRoomTemperatureChange != null)
WhenRoomTemperatureChange(_roomTemperature);
System.Threading.Thread.Sleep(500);//Every second Temperature changes based on AC/Heat Status
}
}
}
public class MySweetHome
{
RoomTemperatureController roomController = null;
public MySweetHome()
{
roomController = new RoomTemperatureController();
roomController.WhenRoomTemperatureChange += TurnHeatOrACBasedOnTemp;
//roomController.WhenRoomTemperatureChange = null; //Setting NULL to delegate reference is possible where as for Event it is not possible.
//roomController.WhenRoomTemperatureChange.DynamicInvoke();//Dynamic Invoke is possible for Delgate and not possible with Event
roomController.SimulateRoomTemperature();
System.Threading.Thread.Sleep(5000);
roomController.TurnAirCondition (true);
roomController.TurnRoomTeperatureSimulator = true;
}
public void TurnHeatOrACBasedOnTemp(int temp)
{
if (temp >= 30)
roomController.TurnAirCondition(true);
if (temp <= 15)
roomController.TurnHeat(true);
}
public static void Main(string []args)
{
MySweetHome home = new MySweetHome();
}
}

For people live in 2020, and want a clean answer...
Definitions:
delegate: defines a function pointer.
event: defines
(1) protected interfaces, and
(2) operations(+=, -=), and
(3) advantage: you don't need to use new keyword anymore.
Regarding the adjective protected:
// eventTest.SomeoneSay = null; // Compile Error.
// eventTest.SomeoneSay = new Say(SayHello); // Compile Error.
Also notice this section from Microsoft: https://learn.microsoft.com/en-us/dotnet/standard/events/#raising-multiple-events
Code Example:
with delegate:
public class DelegateTest
{
public delegate void Say(); // Define a pointer type "void <- ()" named "Say".
private Say say;
public DelegateTest() {
say = new Say(SayHello); // Setup the field, Say say, first.
say += new Say(SayGoodBye);
say.Invoke();
}
public void SayHello() { /* display "Hello World!" to your GUI. */ }
public void SayGoodBye() { /* display "Good bye!" to your GUI. */ }
}
with event:
public class EventTest
{
public delegate void Say();
public event Say SomeoneSay; // Use the type "Say" to define event, an
// auto-setup-everything-good field for you.
public EventTest() {
SomeoneSay += SayHello;
SomeoneSay += SayGoodBye;
SomeoneSay();
}
public void SayHello() { /* display "Hello World!" to your GUI. */ }
public void SayGoodBye() { /* display "Good bye!" to your GUI. */ }
}
Reference:
Event vs. Delegate - Explaining the important differences between the Event and Delegate patterns in C# and why they're useful.: https://dzone.com/articles/event-vs-delegate