I've noticed that my dependency injected, observer-pattern-heavy code (using Guava's EventBus) is often significantly more difficult to debug than code I've written in the past without these features. Particularly when trying to determine when and why observer code is being called.
Martin Oderski and friends wrote a lengthy paper with an especially alluring title, "Deprecating the Observer Pattern" and I have not yet made the time to read it.
I'd like to know what is so wrong with the observer pattern and so much better about the (proposed or other) alternatives to lead such bright people to write this paper.
As a start, I did find one (entertaining) critique of the paper here.
Quoting directly from the paper:
To illustrate the precise problems of the observer pattern,
we start with a simple and ubiquitous example: mouse dragging.
The following example traces the movements of the
mouse during a drag operation in a Path object and displays
it on the screen. To keep things simple, we use Scala closures
as observers.
var path: Path = null
val moveObserver = { (event: MouseEvent) =>
path.lineTo(event.position)
draw(path)
}
control.addMouseDownObserver { event =>
path = new Path(event.position)
control.addMouseMoveObserver(moveObserver)
}
control.addMouseUpObserver { event =>
control.removeMouseMoveObserver(moveObserver)
path.close()
draw(path)
}
The above example, and as we will argue the observer
pattern as defined in [25] in general, violates an impressive
line-up of important software engineering principles:
Side-effects Observers promote side-effects. Since observers
are stateless, we often need several of them to simulate
a state machine as in the drag example. We have to save
the state where it is accessible to all involved observers
such as in the variable path above.
Encapsulation As the state variable path escapes the scope
of the observers, the observer pattern breaks encapsulation.
Composability Multiple observers form a loose collection
of objects that deal with a single concern (or multiple,
see next point). Since multiple observers are installed at
different points at different times, we can’t, for instance,
easily dispose them altogether.
Separation of concerns The above observers not only trace
the mouse path but also call a drawing command, or
more generally, include two different concerns in the
same code location. It is often preferable to separate the
concerns of constructing the path and displaying it, e.g.,
as in the model-view-controller (MVC) [30] pattern.
Scalablity We could achieve a separation of concerns in our
example by creating a class for paths that itself publishes
events when the path changes. Unfortunately, there is no
guarantee for data consistency in the observer pattern.
Let us suppose we would create another event publishing
object that depends on changes in our original path, e.g.,
a rectangle that represents the bounds of our path. Also
consider an observer listening to changes in both the
path and its bounds in order to draw a framed path. This
observer would manually need to determine whether the
bounds are already updated and, if not, defer the drawing
operation. Otherwise the user could observe a frame on
the screen that has the wrong size (a glitch).
Uniformity Different methods to install different observers
decrease code uniformity.
Abstraction There is a low level of abstraction in the example.
It relies on a heavyweight interface of a control
class that provides more than just specific methods to install
mouse event observers. Therefore, we cannot abstract
over the precise event sources. For instance, we
could let the user abort a drag operation by hitting the escape
key or use a different pointer device such as a touch
screen or graphics tablet.
Resource management An observer’s life-time needs to be
managed by clients. Because of performance reasons,
we want to observe mouse move events only during a
drag operation. Therefore, we need to explicitly install
and uninstall the mouse move observer and we need to
remember the point of installation (control above).
Semantic distance Ultimately, the example is hard to understand
because the control flow is inverted which results
in too much boilerplate code that increases the semantic
distance between the programmers intention and
the actual code.
[25] E. Gamma, R. Helm, R. Johnson, and J. Vlissides. Design
patterns: elements of reusable object-oriented software.
Addison-Wesley Longman Publishing Co., Inc., Boston, MA,
USA, 1995. ISBN 0-201-63361-2.
I believe the Observer pattern has the standard drawbacks that come with decoupling things. The Subject gets decoupled from Observer but you cannot just look at its source code and find out who observes it. Hardcoded dependencies are usually easier to read and think about but they are harder to modify and reuse. It's a tradeoff.
As to the paper, it does not address the Observer pattern itself but a particular usage of it. In particular: multiple stateless Observer objects per single object being observed. This has the obvious drawback of the separate observers needing to synchronize with each other ("Since observers are stateless, we often need several of them to simulate
a state machine as in the drag example. We have to save
the state where it is accessible to all involved observers
such as in the variable path above.")
The above drawback is specific to this kind of usage, not to the Observer pattern itself. You could as well create a single (stateful!) observer object that implements all the OnThis, OnThat,OnWhatever methods and get rid of the problem of simulating a state machine across many stateless objects.
I will be brief because I'm new to the topic (and didn't read that specific article yet).
Observer Pattern is intuitively wrong: The Object to be observed knows who is observing (Subject<>--Observer).
That is against real-life (in event-based scenarios). If I scream, I have no idea who is listening; if a lightening, hits the floor... lightning doesn't know that there is a floor till it hits!. Only Observers know what they can observe.
When this kind of of things happen then software use to be a mess -because constructed against our way of thinking-.
It is as if and object knew what other objects can call his methods.
IMO a layer such as "Environment" is the one in charge of taking the events and notifying the affected. (OR mixes the event and the generator of that event)
Event-Source (Subject) generates events to the Environment. Environment delivers the event to the Observer. Observer could register to the kind of events that affects him or it is actually defined in the Environment. Both possibilities make sense (but I wanted to be brief).
In my understanding the Observer Pattern puts together environment & subject.
PS. hate to put in paragraphs abstract ideas! :P
For a 2D game I'm making (for Android) I'm using a component-based system where a GameObject holds several GameComponent objects. GameComponents can be things such as input components, rendering components, bullet emitting components, and so on. Currently, GameComponents have a reference to the object that owns them and can modify it, but the GameObject itself just has a list of components and it doesn't care what the components are as long as they can be updated when the object is updated.
Sometimes a component has some information which the GameObject needs to know. For example, for collision detection a GameObject registers itself with the collision detection subsystem to be notified when it collides with another object. The collision detection subsystem needs to know the object's bounding box. I store x and y in the object directly (because it is used by several components), but width and height are only known to the rendering component which holds the object's bitmap. I would like to have a method getBoundingBox or getWidth in the GameObject that gets that information. Or in general, I want to send some information from a component to the object. However, in my current design the GameObject doesn't know what specific components it has in the list.
I can think of several ways to solve this problem:
Instead of having a completely generic list of components, I can let the GameObject have specific field for some of the important components. For example, it can have a member variable called renderingComponent; whenever I need to get the width of the object I just use renderingComponent.getWidth(). This solution still allows for generic list of components but it treats some of them differently, and I'm afraid I'll end up having several exceptional fields as more components need to be queried. Some objects don't even have rendering components.
Have the required information as members of the GameObject but allow the components to update it. So an object has a width and a height which are 0 or -1 by default, but a rendering component can set them to the correct values in its update loop. This feels like a hack and I might end up pushing many things to the GameObject class for convenience even if not all objects need them.
Have components implement an interface that indicates what type of information they can be queried for. For example, a rendering component would implement the HasSize interface which includes methods such as getWidth and getHeight. When the GameObject needs the width, it loops over its components checking if they implement the HasSize interface (using the instanceof keyword in Java, or is in C#). This seems like a more generic solution, one disadvantage is that searching for the component might take some time (but then, most objects have 3 or 4 components only).
This question isn't about a specific problem. It comes up often in my design and I was wondering what's the best way to handle it. Performance is somewhat important since this is a game, but the number of components per object is generally small (the maximum is 8).
The short version
In a component based system for a game, what is the best way to pass information from the components to the object while keeping the design generic?
We get variations on this question three or four times a week on GameDev.net (where the gameobject is typically called an 'entity') and so far there's no consensus on the best approach. Several different approaches have been shown to be workable however so I wouldn't worry about it too much.
However, usually the problems regard communicating between components. Rarely do people worry about getting information from a component to the entity - if an entity knows what information it needs, then presumably it knows exactly what type of component it needs to access and which property or method it needs to call on that component to get the data. if you need to be reactive rather than active, then register callbacks or have an observer pattern set up with the components to let the entity know when something in the component has changed, and read the value at that point.
Completely generic components are largely useless: they need to provide some sort of known interface otherwise there's little point them existing. Otherwise you may as well just have a large associative array of untyped values and be done with it. In Java, Python, C#, and other slightly-higher-level languages than C++ you can use reflection to give you a more generic way of using specific subclasses without having to encode type and interface information into the components themselves.
As for communication:
Some people are making assumptions that an entity will always contain a known set of component types (where each instance is one of several possible subclasses) and therefore can just grab a direct reference to the other component and read/write via its public interface.
Some people are using publish/subscribe, signals/slots, etc., to create arbitrary connections between components. This seems a bit more flexible but ultimately you still need something with knowledge of these implicit dependencies. (And if this is known at compile time, why not just use the previous approach?)
Or, you can put all shared data in the entity itself and use that as a shared communication area (tenuously related to the blackboard system in AI) that each of the components can read and write to. This usually requires some robustness in the face of certain properties not existing when you expected them to. It also doesn't lend itself to parallelism, although I doubt that's a massive concern on a small embedded system...?
Finally, some people have systems where the entity doesn't exist at all. The components live within their subsystems and the only notion of an entity is an ID value in certain components - if a Rendering component (within the Rendering system) and a Player component (within the Players system) have the same ID, then you can assume the former handles the drawing of the latter. But there isn't any single object that aggregates either of those components.
Like others have said, there's no always right answer here. Different games will lend themselves towards different solutions. If you're building a big complex game with lots of different kinds of entities, a more decoupled generic architecture with some kind of abstract messaging between components may be worth the effort for the maintainability you get. For a simpler game with similar entities, it may make the most sense to just push all of that state up into GameObject.
For your specific scenario where you need to store the bounding box somewhere and only the collision component cares about it, I would:
Store it in the collision component itself.
Make the collision detection code work with the components directly.
So, instead of having the collision engine iterate through a collection of GameObjects to resolve the interaction, have it iterate directly through a collection of CollisionComponents. Once a collision has occurred, it will be up to the component to push that up to its parent GameObject.
This gives you a couple of benefits:
Leaves collision-specific state out of GameObject.
Spares you from iterating over GameObjects that don't have collision components. (If you have a lot of non-interactive objects like visual effects and decoration, this can save a decent number of cycles.)
Spares you from burning cycles walking between the object and its component. If you iterate through the objects then do getCollisionComponent() on each one, that pointer-following can cause a cache miss. Doing that for every frame for every object can burn a lot of CPU.
If you're interested I have more on this pattern here, although it looks like you already understand most of what's in that chapter.
Use an "event bus". (note that you probably can't use the code as is but it should give you the basic idea).
Basically, create a central resource where every object can register itself as a listener and say "If X happens, I want to know". When something happens in the game, the responsible object can simply send an event X to the event bus and all interesting parties will notice.
[EDIT] For a more detailed discussion, see message passing (thanks to snk_kid for pointing this out).
One approach is to initialize a container of components. Each component can provide a service and may also require services from other components. Depending on your programming language and environment you have to come up with a method for providing this information.
In its simplest form you have one-to-one connections between components, but you will also need one-to-many connections. E.g. the CollectionDetector will have a list of components implementing IBoundingBox.
During initialization the container will wire up connections between components, and during run-time there will be no additional cost.
This is close to you solution 3), expect the connections between components are wired only once and are not checked at every iteration of the game loop.
The Managed Extensibility Framework for .NET is a nice solution to this problem. I realize that you intend to develop on Android, but you may still get some inspiration from this framework.
Inversion of Control (IoC) can be quite confusing when it is first encountered.
What is it?
Which problem does it solve?
When is it appropriate to use and when not?
The Inversion-of-Control (IoC) pattern, is about providing any kind of callback, which "implements" and/or controls reaction, instead of acting ourselves directly (in other words, inversion and/or redirecting control to the external handler/controller).
The Dependency-Injection (DI) pattern is a more specific version of IoC pattern, and is all about removing dependencies from your code.
Every DI implementation can be considered IoC, but one should not call it IoC, because implementing Dependency-Injection is harder than callback (Don't lower your product's worth by using the general term "IoC" instead).
For DI example, say your application has a text-editor component, and you want to provide spell checking. Your standard code would look something like this:
public class TextEditor {
private SpellChecker checker;
public TextEditor() {
this.checker = new SpellChecker();
}
}
What we've done here creates a dependency between the TextEditor and the SpellChecker.
In an IoC scenario we would instead do something like this:
public class TextEditor {
private IocSpellChecker checker;
public TextEditor(IocSpellChecker checker) {
this.checker = checker;
}
}
In the first code example we are instantiating SpellChecker (this.checker = new SpellChecker();), which means the TextEditor class directly depends on the SpellChecker class.
In the second code example we are creating an abstraction by having the SpellChecker dependency class in TextEditor's constructor signature (not initializing dependency in class). This allows us to call the dependency then pass it to the TextEditor class like so:
SpellChecker sc = new SpellChecker(); // dependency
TextEditor textEditor = new TextEditor(sc);
Now the client creating the TextEditor class has control over which SpellChecker implementation to use because we're injecting the dependency into the TextEditor signature.
Note that just like IoC being the base of many other patterns, above sample is only one of many Dependency-Injection kinds, for example:
Constructor Injection.
Where an instance of IocSpellChecker would be passed to constructor, either automatically or similar to above manually.
Setter Injection.
Where an instance of IocSpellChecker would be passed through setter-method or public property.
Service-lookup and/or Service-locator
Where TextEditor would ask a known provider for a globally-used-instance (service) of IocSpellChecker type (and that maybe without storing said instance, and instead, asking the provider again and again).
Inversion of Control is what you get when your program callbacks, e.g. like a gui program.
For example, in an old school menu, you might have:
print "enter your name"
read name
print "enter your address"
read address
etc...
store in database
thereby controlling the flow of user interaction.
In a GUI program or somesuch, instead we say:
when the user types in field a, store it in NAME
when the user types in field b, store it in ADDRESS
when the user clicks the save button, call StoreInDatabase
So now control is inverted... instead of the computer accepting user input in a fixed order, the user controls the order in which the data is entered, and when the data is saved in the database.
Basically, anything with an event loop, callbacks, or execute triggers falls into this category.
What is Inversion of Control?
If you follow these simple two steps, you have done inversion of control:
Separate what-to-do part from when-to-do part.
Ensure that when part knows as little as possible about what part; and vice versa.
There are several techniques possible for each of these steps based on the technology/language you are using for your implementation.
--
The inversion part of the Inversion of Control (IoC) is the confusing thing; because inversion is the relative term. The best way to understand IoC is to forget about that word!
--
Examples
Event Handling. Event Handlers (what-to-do part) -- Raising Events (when-to-do part)
Dependency Injection. Code that constructs a dependency (what-to-do part) -- instantiating and injecting that dependency for the clients when needed, which is usually taken care of by the DI tools such as Dagger (when-to-do-part).
Interfaces. Component client (when-to-do part) -- Component Interface implementation (what-to-do part)
xUnit fixture. Setup and TearDown (what-to-do part) -- xUnit frameworks calls to Setup at the beginning and TearDown at the end (when-to-do part)
Template method design pattern. template method when-to-do part -- primitive subclass implementation what-to-do part
DLL container methods in COM. DllMain, DllCanUnload, etc (what-to-do part) -- COM/OS (when-to-do part)
Inversion of Controls is about separating concerns.
Without IoC: You have a laptop computer and you accidentally break the screen. And darn, you find the same model laptop screen is nowhere in the market. So you're stuck.
With IoC: You have a desktop computer and you accidentally break the screen. You find you can just grab almost any desktop monitor from the market, and it works well with your desktop.
Your desktop successfully implements IoC in this case. It accepts a variety type of monitors, while the laptop does not, it needs a specific screen to get fixed.
Inversion of Control, (or IoC), is about getting freedom (You get married, you lost freedom and you are being controlled. You divorced, you have just implemented Inversion of Control. That's what we called, "decoupled". Good computer system discourages some very close relationship.) more flexibility (The kitchen in your office only serves clean tap water, that is your only choice when you want to drink. Your boss implemented Inversion of Control by setting up a new coffee machine. Now you get the flexibility of choosing either tap water or coffee.) and less dependency (Your partner has a job, you don't have a job, you financially depend on your partner, so you are controlled. You find a job, you have implemented Inversion of Control. Good computer system encourages in-dependency.)
When you use a desktop computer, you have slaved (or say, controlled). You have to sit before a screen and look at it. Using the keyboard to type and using the mouse to navigate. And a badly written software can slave you even more. If you replace your desktop with a laptop, then you somewhat inverted control. You can easily take it and move around. So now you can control where you are with your computer, instead of your computer controlling it.
By implementing Inversion of Control, a software/object consumer gets more controls/options over the software/objects, instead of being controlled or having fewer options.
With the above ideas in mind. We still miss a key part of IoC. In the scenario of IoC, the software/object consumer is a sophisticated framework. That means the code you created is not called by yourself. Now let's explain why this way works better for a web application.
Suppose your code is a group of workers. They need to build a car. These workers need a place and tools (a software framework) to build the car. A traditional software framework will be like a garage with many tools. So the workers need to make a plan themselves and use the tools to build the car. Building a car is not an easy business, it will be really hard for the workers to plan and cooperate properly. A modern software framework will be like a modern car factory with all the facilities and managers in place. The workers do not have to make any plan, the managers (part of the framework, they are the smartest people and made the most sophisticated plan) will help coordinate so that the workers know when to do their job (framework calls your code). The workers just need to be flexible enough to use any tools the managers give to them (by using Dependency Injection).
Although the workers give the control of managing the project on the top level to the managers (the framework). But it is good to have some professionals help out. This is the concept of IoC truly come from.
Modern Web applications with an MVC architecture depends on the framework to do URL Routing and put Controllers in place for the framework to call.
Dependency Injection and Inversion of Control are related. Dependency Injection is at the micro level and Inversion of Control is at the macro level. You have to eat every bite (implement DI) in order to finish a meal (implement IoC).
Before using Inversion of Control you should be well aware of the fact that it has its pros and cons and you should know why you use it if you do so.
Pros:
Your code gets decoupled so you can easily exchange implementations of an interface with alternative implementations
It is a strong motivator for coding against interfaces instead of implementations
It's very easy to write unit tests for your code because it depends on nothing else than the objects it accepts in its constructor/setters and you can easily initialize them with the right objects in isolation.
Cons:
IoC not only inverts the control flow in your program, it also clouds it considerably. This means you can no longer just read your code and jump from one place to another because the connections that would normally be in your code are not in the code anymore. Instead it is in XML configuration files or annotations and in the code of your IoC container that interprets these metadata.
There arises a new class of bugs where you get your XML config or your annotations wrong and you can spend a lot of time finding out why your IoC container injects a null reference into one of your objects under certain conditions.
Personally I see the strong points of IoC and I really like them but I tend to avoid IoC whenever possible because it turns your software into a collection of classes that no longer constitute a "real" program but just something that needs to be put together by XML configuration or annotation metadata and would fall (and falls) apart without it.
Wikipedia Article. To me, inversion of control is turning your sequentially written code and turning it into an delegation structure. Instead of your program explicitly controlling everything, your program sets up a class or library with certain functions to be called when certain things happen.
It solves code duplication. For example, in the old days you would manually write your own event loop, polling the system libraries for new events. Nowadays, most modern APIs you simply tell the system libraries what events you're interested in, and it will let you know when they happen.
Inversion of control is a practical way to reduce code duplication, and if you find yourself copying an entire method and only changing a small piece of the code, you can consider tackling it with inversion of control. Inversion of control is made easy in many languages through the concept of delegates, interfaces, or even raw function pointers.
It is not appropriate to use in all cases, because the flow of a program can be harder to follow when written this way. It's a useful way to design methods when writing a library that will be reused, but it should be used sparingly in the core of your own program unless it really solves a code duplication problem.
Suppose you are an object. And you go to a restaurant:
Without IoC: you ask for "apple", and you are always served apple when you ask more.
With IoC: You can ask for "fruit". You can get different fruits each time you get served. for example, apple, orange, or water melon.
So, obviously, IoC is preferred when you like the varieties.
Answering only the first part.
What is it?
Inversion of Control (IoC) means to create instances of dependencies first and latter instance of a class (optionally injecting them through constructor), instead of creating an instance of the class first and then the class instance creating instances of dependencies.
Thus, inversion of control inverts the flow of control of the program. Instead of the callee controlling the flow of control (while creating dependencies), the caller controls the flow of control of the program.
But I think you have to be very careful with it. If you will overuse this pattern, you will make very complicated design and even more complicated code.
Like in this example with TextEditor: if you have only one SpellChecker maybe it is not really necessary to use IoC ? Unless you need to write unit tests or something ...
Anyway: be reasonable. Design pattern are good practices but not Bible to be preached. Do not stick it everywhere.
IoC / DI to me is pushing out dependencies to the calling objects. Super simple.
The non-techy answer is being able to swap out an engine in a car right before you turn it on. If everything hooks up right (the interface), you are good.
Inversion of control is a pattern used for decoupling components and layers in the system. The pattern is implemented through injecting dependencies into a component when it is constructed. These dependences are usually provided as interfaces for further decoupling and to support testability. IoC / DI containers such as Castle Windsor, Unity are tools (libraries) which can be used for providing IoC. These tools provide extended features above and beyond simple dependency management, including lifetime, AOP / Interception, policy, etc.
a. Alleviates a component from being responsible for managing it's dependencies.
b. Provides the ability to swap dependency implementations in different environments.
c. Allows a component be tested through mocking of dependencies.
d. Provides a mechanism for sharing resources throughout an application.
a. Critical when doing test-driven development. Without IoC it can be difficult to test, because the components under test are highly coupled to the rest of the system.
b. Critical when developing modular systems. A modular system is a system whose components can be replaced without requiring recompilation.
c. Critical if there are many cross-cutting concerns which need to addressed, partilarly in an enterprise application.
Let's say that we have a meeting in a hotel.
We have invited many people, so we have left out many jugs of water and many plastic cups.
When somebody wants to drink, he/she fills a cup, drinks the water and throws the cup on the floor.
After an hour or so we have a floor covered with plastic cups and water.
Let's try that after inverting the control:
Imagine the same meeting in the same place, but instead of plastic cups we now have a waiter with just one glass cup (Singleton)
When somebody wants to drink, the waiter gets one for them. They drink it and return it to the waiter.
Leaving aside the question of the hygiene, the use of a waiter (process control) is much more effective and economic.
And this is exactly what Spring (another IoC container, for example: Guice) does. Instead of letting the application create what it needs using the new keyword (i.e. taking a plastic cup), Spring IoC offers the application the same cup/ instance (singleton) of the needed object (glass of water).
Think of yourself as an organizer of such a meeting:
Example:-
public class MeetingMember {
private GlassOfWater glassOfWater;
...
public void setGlassOfWater(GlassOfWater glassOfWater){
this.glassOfWater = glassOfWater;
}
//your glassOfWater object initialized and ready to use...
//spring IoC called setGlassOfWater method itself in order to
//offer to meetingMember glassOfWater instance
}
Useful links:-
http://adfjsf.blogspot.in/2008/05/inversion-of-control.html
http://martinfowler.com/articles/injection.html
http://www.shawn-barrett.com/blog/post/Tip-of-the-day-e28093-Inversion-Of-Control.aspx
I shall write down my simple understanding of this two terms:
For quick understanding just read examples*
Dependency Injection(DI):
Dependency injection generally means passing an object on which method depends, as a parameter to a method, rather than having the method create the dependent object. What it means in practice is that the method does not depends directly on a particular implementation; any implementation that meets the requirements can be passed as a parameter.
With this objects tell thier dependencies.
And spring makes it available. This leads to loosely coupled application development.
Quick Example:EMPLOYEE OBJECT WHEN CREATED,
IT WILL AUTOMATICALLY CREATE ADDRESS OBJECT
(if address is defines as dependency by Employee object)
Inversion of Control(IoC) Container:
This is common characteristic of frameworks,
IOC manages java objects – from instantiation to destruction through its BeanFactory. -Java components that are instantiated by the IoC container are called beans, and the IoC container manages a bean's scope, lifecycle events, and any AOP features for which it has been configured and coded.
QUICK EXAMPLE:Inversion of Control is about getting freedom, more flexibility, and less dependency. When you are using a desktop computer, you are slaved (or say, controlled). You have to sit before a screen and look at it. Using keyboard to type and using mouse to navigate. And a bad written software can slave you even more. If you replaced your desktop with a laptop, then you somewhat inverted control. You can easily take it and move around. So now you can control where you are with your computer, instead of computer controlling it.
By implementing Inversion of Control, a software/object consumer get more controls/options over the software/objects, instead of being controlled or having less options.
Inversion of control as a design guideline serves the following purposes:
There is a decoupling of the execution of a certain task from implementation.
Every module can focus on what it is designed for.
Modules make no assumptions about what other systems do but rely on their contracts.
Replacing modules has no side effect on other modules I will keep things abstract here, You can visit following links for detail understanding of the topic.
A good read with example
Detailed explanation
I found a very clear example here which explains how the 'control is inverted'.
Classic code (without Dependency injection)
Here is how a code not using DI will roughly work:
Application needs Foo (e.g. a controller), so:
Application creates Foo
Application calls Foo
Foo needs Bar (e.g. a service), so:
Foo creates Bar
Foo calls Bar
Bar needs Bim (a service, a repository, …), so:
Bar creates Bim
Bar does something
Using dependency injection
Here is how a code using DI will roughly work:
Application needs Foo, which needs Bar, which needs Bim, so:
Application creates Bim
Application creates Bar and gives it Bim
Application creates Foo and gives it Bar
Application calls Foo
Foo calls Bar
Bar does something
The control of the dependencies is inverted from one being called to the one calling.
What problems does it solve?
Dependency injection makes it easy to swap with the different implementation of the injected classes. While unit testing you can inject a dummy implementation, which makes the testing a lot easier.
Ex: Suppose your application stores the user uploaded file in the Google Drive, with DI your controller code may look like this:
class SomeController
{
private $storage;
function __construct(StorageServiceInterface $storage)
{
$this->storage = $storage;
}
public function myFunction ()
{
return $this->storage->getFile($fileName);
}
}
class GoogleDriveService implements StorageServiceInterface
{
public function authenticate($user) {}
public function putFile($file) {}
public function getFile($file) {}
}
When your requirements change say, instead of GoogleDrive you are asked to use the Dropbox. You only need to write a dropbox implementation for the StorageServiceInterface. You don't have make any changes in the controller as long as Dropbox implementation adheres to the StorageServiceInterface.
While testing you can create the mock for the StorageServiceInterface with the dummy implementation where all the methods return null(or any predefined value as per your testing requirement).
Instead if you had the controller class to construct the storage object with the new keyword like this:
class SomeController
{
private $storage;
function __construct()
{
$this->storage = new GoogleDriveService();
}
public function myFunction ()
{
return $this->storage->getFile($fileName);
}
}
When you want to change with the Dropbox implementation you have to replace all the lines where new GoogleDriveService object is constructed and use the DropboxService. Besides when testing the SomeController class the constructor always expects the GoogleDriveService class and the actual methods of this class are triggered.
When is it appropriate and when not?
In my opinion you use DI when you think there are (or there can be) alternative implementations of a class.
I agree with NilObject, but I'd like to add to this:
if you find yourself copying an entire method and only changing a small piece of the code, you can consider tackling it with inversion of control
If you find yourself copying and pasting code around, you're almost always doing something wrong. Codified as the design principle Once and Only Once.
For example, task#1 is to create object.
Without IOC concept, task#1 is supposed to be done by Programmer.But With IOC concept, task#1 would be done by container.
In short Control gets inverted from Programmer to container. So, it is called as inversion of control.
I found one good example here.
It seems that the most confusing thing about "IoC" the acronym and the name for which it stands is that it's too glamorous of a name - almost a noise name.
Do we really need a name by which to describe the difference between procedural and event driven programming? OK, if we need to, but do we need to pick a brand new "bigger than life" name that confuses more than it solves?
Inversion of control is when you go to the grocery store and your wife gives you the list of products to buy.
In programming terms, she passed a callback function getProductList() to the function you are executing - doShopping().
It allows user of the function to define some parts of it, making it more flexible.
I understand that the answer has already been given here. But I still think, some basics about the inversion of control have to be discussed here in length for future readers.
Inversion of Control (IoC) has been built on a very simple principle called Hollywood Principle. And it says that,
Don't call us, we'll call you
What it means is that don't go to the Hollywood to fulfill your dream rather if you are worthy then Hollywood will find you and make your dream comes true. Pretty much inverted, huh?
Now when we discuss about the principle of IoC, we use to forget about the Hollywood. For IoC, there has to be three element, a Hollywood, you and a task like to fulfill your dream.
In our programming world, Hollywood represent a generic framework (may be written by you or someone else), you represent the user code you wrote and the task represent the thing you want to accomplish with your code. Now you don't ever go to trigger your task by yourself, not in IoC! Rather you have designed everything in such that your framework will trigger your task for you. Thus you have built a reusable framework which can make someone a hero or another one a villain. But that framework is always in charge, it knows when to pick someone and that someone only knows what it wants to be.
A real life example would be given here. Suppose, you want to develop a web application. So, you create a framework which will handle all the common things a web application should handle like handling http request, creating application menu, serving pages, managing cookies, triggering events etc.
And then you leave some hooks in your framework where you can put further codes to generate custom menu, pages, cookies or logging some user events etc. On every browser request, your framework will run and executes your custom codes if hooked then serve it back to the browser.
So, the idea is pretty much simple. Rather than creating a user application which will control everything, first you create a reusable framework which will control everything then write your custom codes and hook it to the framework to execute those in time.
Laravel and EJB are examples of such a frameworks.
Reference:
https://martinfowler.com/bliki/InversionOfControl.html
https://en.wikipedia.org/wiki/Inversion_of_control
Inversion of Control is a generic principle, while Dependency Injection realises this principle as a design pattern for object graph construction (i.e. configuration controls how the objects are referencing each other, rather than the object itself controlling how to get the reference to another object).
Looking at Inversion of Control as a design pattern, we need to look at what we are inverting. Dependency Injection inverts control of constructing a graph of objects. If told in layman's term, inversion of control implies change in flow of control in the program. Eg. In traditional standalone app, we have main method, from where the control gets passed to other third party libraries(in case, we have used third party library's function), but through inversion of control control gets transferred from third party library code to our code, as we are taking the service of third party library. But there are other aspects that need to be inverted within a program - e.g. invocation of methods and threads to execute the code.
For those interested in more depth on Inversion of Control a paper has been published outlining a more complete picture of Inversion of Control as a design pattern (OfficeFloor: using office patterns to improve software design http://doi.acm.org/10.1145/2739011.2739013 with a free copy available to download from http://www.officefloor.net/about.html).
What is identified is the following relationship:
Inversion of Control (for methods) = Dependency (state) Injection + Continuation Injection + Thread Injection
Summary of above relationship for Inversion of Control available - http://dzone.com/articles/inversion-of-coupling-control
IoC is about inverting the relationship between your code and third-party code (library/framework):
In normal s/w development, you write the main() method and call "library" methods. You are in control :)
In IoC the "framework" controls main() and calls your methods. The Framework is in control :(
DI (Dependency Injection) is about how the control flows in the application. Traditional desktop application had control flow from your application(main() method) to other library method calls, but with DI control flow is inverted that's framework takes care of starting your app, initializing it and invoking your methods whenever required.
In the end you always win :)
I like this explanation: http://joelabrahamsson.com/inversion-of-control-an-introduction-with-examples-in-net/
It start simple and shows code examples as well.
The consumer, X, needs the consumed class, Y, to accomplish something. That’s all good and natural, but does X really need to know that it uses Y?
Isn’t it enough that X knows that it uses something that has the behavior, the methods, properties etc, of Y without knowing who actually implements the behavior?
By extracting an abstract definition of the behavior used by X in Y, illustrated as I below, and letting the consumer X use an instance of that instead of Y it can continue to do what it does without having to know the specifics about Y.
In the illustration above Y implements I and X uses an instance of I. While it’s quite possible that X still uses Y what’s interesting is that X doesn’t know that. It just knows that it uses something that implements I.
Read article for further info and description of benefits such as:
X is not dependent on Y anymore
More flexible, implementation can be decided in runtime
Isolation of code unit, easier testing
...
A very simple written explanation can be found here
http://binstock.blogspot.in/2008/01/excellent-explanation-of-dependency.html
It says -
"Any nontrivial application is made up of two or more classes that
collaborate with each other to perform some business logic.
Traditionally, each object is responsible for obtaining its own
references to the objects it collaborates with (its dependencies).
When applying DI, the objects are given their dependencies at creation
time by some external entity that coordinates each object in the
system. In other words, dependencies are injected into objects."
Programming speaking
IoC in easy terms: It's the use of Interface as a way of specific something (such a field or a parameter) as a wildcard that can be used by some classes. It allows the re-usability of the code.
For example, let's say that we have two classes : Dog and Cat. Both shares the same qualities/states: age, size, weight. So instead of creating a class of service called DogService and CatService, I can create a single one called AnimalService that allows to use Dog and Cat only if they use the interface IAnimal.
However, pragmatically speaking, it has some backwards.
a) Most of the developers don't know how to use it. For example, I can create a class called Customer and I can create automatically (using the tools of the IDE) an interface called ICustomer. So, it's not rare to find a folder filled with classes and interfaces, no matter if the interfaces will be reused or not. It's called BLOATED. Some people could argue that "may be in the future we could use it". :-|
b) It has some limitings. For example, let's talk about the case of Dog and Cat and I want to add a new service (functionality) only for dogs. Let's say that I want to calculate the number of days that I need to train a dog (trainDays()), for cat it's useless, cats can't be trained (I'm joking).
b.1) If I add trainDays() to the Service AnimalService then it also works with cats and it's not valid at all.
b.2) I can add a condition in trainDays() where it evaluates which class is used. But it will break completely the IoC.
b.3) I can create a new class of service called DogService just for the new functionality. But, it will increase the maintainability of the code because we will have two classes of service (with similar functionality) for Dog and it's bad.
Inversion of control is about transferring control from library to the client. It makes more sense when we talk about a client that injects (passes) a function value (lambda expression) into a higher order function (library function) that controls (changes) the behavior of the library function.
So, a simple implementation (with huge implications) of this pattern is a higher order library function (which accepts another function as an argument). The library function transfers control over its behavior by giving the client the ability to supply the "control" function as an argument.
For example, library functions like "map", "flatMap" are IoC implementations.
Of course, a limited IoC version is, for example, a boolean function parameter. A client may control the library function by switching the boolean argument.
A client or framework that injects library dependencies (which carry behavior) into libraries may also be considered IoC
I've read a lot of answers for this but if someone is still confused and needs a plus ultra "laymans term" to explain IoC here is my take:
Imagine a parent and child talking to each other.
Without IoC:
*Parent: You can only speak when I ask you questions and you can only act when I give you permission.
Parent: This means, you can't ask me if you can eat, play, go to the bathroom or even sleep if I don't ask you.
Parent: Do you want to eat?
Child: No.
Parent: Okay, I'll be back. Wait for me.
Child: (Wants to play but since there's no question from the parent, the child can't do anything).
After 1 hour...
Parent: I'm back. Do you want to play?
Child: Yes.
Parent: Permission granted.
Child: (finally is able to play).
This simple scenario explains the control is centered to the parent. The child's freedom is restricted and highly depends on the parent's question. The child can ONLY speak when asked to speak, and can ONLY act when granted permission.
With IoC:
The child has now the ability to ask questions and the parent can respond with answers and permissions. Simply means the control is inverted!
The child is now free to ask questions anytime and though there is still dependency with the parent regarding permissions, he is not dependent in the means of speaking/asking questions.
In a technological way of explaining, this is very similar to console/shell/cmd vs GUI interaction. (Which is answer of Mark Harrison above no.2 top answer).
In console, you are dependent on the what is being asked/displayed to you and you can't jump to other menus and features without answering it's question first; following a strict sequential flow. (programmatically this is like a method/function loop).
However with GUI, the menus and features are laid out and the user can select whatever it needs thus having more control and being less restricted. (programmatically, menus have callback when selected and an action takes place).
Since already there are many answers for the question but none of them shows the breakdown of Inversion Control term I see an opportunity to give a more concise and useful answer.
Inversion of Control is a pattern that implements the Dependency Inversion Principle (DIP). DIP states the following: 1. High-level modules should not depend on low-level modules. Both should depend on abstractions (e.g. interfaces). 2. Abstractions should not depend on details. Details (concrete implementations) should depend on abstractions.
There are three types of Inversion of Control:
Interface Inversion
Providers shouldn’t define an interface. Instead, the consumer should define the interface and providers must implement it. Interface Inversion allows eliminating the necessity to modify the consumer each time when a new provider added.
Flow Inversion
Changes control of the flow. For example, you have a console application where you asked to enter many parameters and after each entered parameter you are forced to press Enter. You can apply Flow Inversion here and implement a desktop application where the user can choose the sequence of parameters’ entering, the user can edit parameters, and at the final step, the user needs to press Enter only once.
Creation Inversion
It can be implemented by the following patterns: Factory Pattern, Service Locator, and Dependency Injection. Creation Inversion helps to eliminate dependencies between types moving the process of dependency objects creation outside of the type that uses these dependency objects. Why dependencies are bad? Here are a couple of examples: direct creation of a new object in your code makes testing harder; it is impossible to change references in assemblies without recompilation (OCP principle violation); you can’t easily replace a desktop-UI by a web-UI.
Creating an object within class is called tight coupling, Spring removes this dependency by following a design pattern(DI/IOC). In which object of class in passed in constructor rather than creating in class. More over we give super class reference variable in constructor to define more general structure.
Using IoC you are not new'ing up your objects. Your IoC container will do that and manage the lifetime of them.
It solves the problem of having to manually change every instantiation of one type of object to another.
It is appropriate when you have functionality that may change in the future or that may be different depending on the environment or configuration used in.