I'm going through the docs because I am about to implement a protocol instead of a class (something I've never done before), and I'm curious as to the difference between the two.
Can someone give an example in plain words?
Thanks
A class serves as a blueprint for creating one or more objects based on specific implementation of that class.
A good analogy is a form for cutting out butter-cookies. The form‘s attributes (shape, size, height) define the cookies that you can cut out with it. You have only one form (class) but you can create many cookies (instances of that class, ie. objects) with it. All cookies are based on that particular form.
Similarily all objects that are instances of that class are identical in their attributes.
Classes = data and methods (special functions), all sophistically bundled together.
Classes define, what their inner content (data) is + what kind of work (methods) they can do.
The content is based on variables that hold various number types, strings, constants, and other more sophisiticated content + methods which are chunks of code that (when executed) perform some computational operations with various data.
All methods defined in class have their
Definition - that defines the name of the method + what (if any) data the methods takes in for processing and what (if any) data the methods spits out for processing by someone else. All methods defined in class also have Implementation – the actual code that provides the processing – it is the innerworkings of methods.. inside there is code that processes the data and also that is able to ask other methods for subprocessing data. So the class is a very noble type in programming.
If you understand the above, you will understand what a protocol is.
A protocol is a set of one or more method declarations and that set has a name and represents a protocol. I say declarations, because the methods that together are defined by a particular protocol, do not have any implementation code defined.. The only thing that exist is their names declared.
Look above - in class, you have always defined not only what methods the class has, but also how that work will be done. But methods in protocol do not have any implementation.
Lets have a real life analogy again, it helps. If you come to my house to live here for a week, you will need to adhere to my TidyUp protocol. The TidyUp protocol defines three methods - wash the dishes every day, clean the room, and ventilate fresh air. These three methods, I define them..are something you will do. But I absolutely do not care, how the implementation should look like, I just nominaly define the methods. You will implement them, ie.you define how the details of that work (those methods) will look like. I just say, adhere to my protocol and implement it as you see fit.
Finale – You can declare some class. You can separately also declare a protocol. And you can then declare, that this class, in addition to its own methods, will adopt or adhere to that protocol, ie. the class wil implement the protocol’s methods.
The plain words from The Objective-C Programming Language explain the purpose of protocols simply:
Protocols declare methods that can be implemented by any class.
Protocols are useful in at least three situations:
To declare methods that others are expected to implement
To declare the interface to an object while concealing its class
To capture similarities among classes that are not hierarchically related
So, protocols declare methods, but don't provide the implementation. A class that adopts a protocol is expected to implement the protocol's methods.
Delegation is a good example of why a protocol is useful. Consider, for example, the UITableViewDataSource protocol. Any class can adopt that protocol, and any class that does so can be used as the data source for a table. A table view doesn't care what kind of object is acting as its data source; it only cares that the object acting as data source implements a particular set of methods. You could use inheritance for this, but then all data source objects would have to be derived from a common base class (more specific than NSObject). Using the protocol instead lets the table count on being able to call methods like -tableView:willBeginEditingRowAtIndexPath: and -tableView:heightForRowAtIndexPath: without needing to know anything else about the data source.
A protocol is a lot like an interface in Java and other languages. Think of it as a contract that describes the interface other classes agree to implement. It can define a list of required and optional methods that an implementing class will implement. Unlike a class, it does not provide its own implementations of those methods.
the main difference between classes and protocols is that writing protocols is useful to implement delegate methods.
in example we've got class A and class B and we want to call a method in class A from the class B.
you can read a very valuable example of that in this article
http://iosdevelopertips.com/objective-c/the-basics-of-protocols-and-delegates.html
reading code is worth a thousand words ;-)
that helped me out the first time I had to use'em
Somewhat less difference than in other languages. An interface (equivalent to a Java/C++ class) defines the data layout of objects and may define some subset of their methods (including the possibility of defining the entire set, of course). A protocol defines a subset of methods only, with no data definition.
Of significance is that a interface can inherit from only one other interface (which can, of course, inherit from an interface which inherits from an interface which inherits ...), but an interface can implement any number of protocols. So two distinct interfaces with no common inheritance (other than NSObject) can both implement the same protocol and thus "certify" that they provide the same functions. (Though with Objective-C you can, with a few tricks, call methods of an interface that aren't externally declared in either the interface declaration or a protocol, so protocols are to a degree just "syntactic sugar" or some such.)
Protocol defines what a class could do, like a Interface in Java or c#
A class is the actual implementation that does the job.
Simple enough? :)
Related
Is there an Interface that I can extend or some other way to create an Interface whereby the implementing class must be a data class? It would be useful to have access to the data class API methods such as copy().
No, copy method have unique number of parameters for every data class, so it's useless to have such interface. If all your data classes have same field - just create and implement common interface.
So I'm going to preface my answer by saying I don't have experience with Kotlin, but I have plenty of Java experience which as I understand it is similar, so unless Kotlin has a feature that helps do what you want that Java doesn't, my answer might still apply.
If I understand correctly, basically what you're trying to do is enforce that whatever class implements your interface X, must also be a subtype of Y.
My first question would be Why would you want to do this? Enforcing that X only be implemented by subtypes of Y is mixing interface and implementation, which the exact opposite of what interfaces are for.
To even enforce this, you would have to have X extend the interface of Y, either implicitly or explicitly. Since in Java (and presumably Kotlin), interfaces cannot extend objects, you have two options:
1) extend the INTERFACE of data, if it exists (which I don't think it does given what I've been reading about data classes. It sounds more like a baked in language feature than just a helpful code object)
2) Add to your interface the exact method signatures of the methods you want out of data classes. BY doing this, you've gained two things: First, you get your convenience methods whenever a data class implements your interface, and second, you still have the flexibility that interfaces are meant to provide, because now even non-data classes can implement your interface if you need them to, they just have to be sure to define the data classes interface methods manually.
This question already has answers here:
What is a Protocol?
(2 answers)
Closed 10 years ago.
Why would I want to use a protocol rather than create a subclass and inherit the methods..?
Please explain it to me, i'm to confused about this topic, i'm not very pleased with the explanation in the book im reading.
Where do I use protocols instead of other ways to get the methods..? if I can subclass a class and get the methods why would i want to use a protocol where i need to define the methods?
Why would I want to use a protocol rather than create a subclass and
inherit the methods..?
Protocols make it possible for unrelated classes to all implement the same interface. Instances of each of those classes can then be used by a client of the protocol. For example, UITableViewDataSource is a protocol that provides an interface by which a table can ask for data from any object that implements the protocol. The table view doesn't care what the type of the object is so long as it implements the data source interface.
Imagine how unpleasant things would be if all table data sources had to inherit from a common class! Objective-C only provides single inheritance, so you'd effectively be constrained to using only a single kind of object for your data source. With protocols, though, a data source can be a view controller, a model object, or perhaps even a remote object.
To be more specific, protocols allow a form of polymorphism. That means that a single object can take several forms: e.g. view controller, table data source, table delegate, scroll view delegate. Because Objective-C is a single-inheritance language, you only get one of those interfaces via inheritance. The rest you implement yourself, but that often makes sense because you generally adopt a given protocol in order to customize some other object's behavior anyway.
Because subclassing and protocols are two different things. Subclassing extends a class with new functionality while inheriting all previous functionality of a specific class while a protocol, when applied to a class, simply adds functionality to it and doesn't inherit anything from it; what that class is usually doesn't matter.
Protocols are most frequently used for the delegate pattern in Objective-C whereby an object can send a message to another object without caring WHAT that object is (i.e. it's class).
Often times, a delegate is declared as:
#property(nonatomic, assign) id < MyObjectDelegate > delegate;
Notice that the class of the property is id -- in essence, you don't care if the object is a car or a turtle -- all you need to know is that it is an object (id) and that it contractually subscribes to the functions you need it to. So if your delegate is type turtle, you can call [delegate myStateChanged]; or, if your delegate is a car, you can call [delegate myStateChanged]. All you need to know is that, if you send a message to it, it will accept it.
I would look up and read about the use of Objective-C delegates as I think it will really help you understand protocols better and how it's different than subclassing. I don't know if you're familiar with other, object-oriented programming languages, but if so, protocols are most similar to Interfaces in other languages.
Protocols are useful because you can implement many protocols, instead you can only extend a single class.
I was just wondering are there other usages of a protocol other than a delegate? I've only seen protocol's been used as delegates but I haven't noticed if it has any other purpose. So are there actual purposes of protocol's other than a delegate?
From the apple docs:
http://developer.apple.com/library/mac/#documentation/Cocoa/Conceptual/ObjectiveC/Chapters/ocProtocols.html
Protocols declare methods that can be implemented by any class. Protocols are useful in at least three situations:
To declare methods that others are expected to implement
To declare the interface to an object while concealing its class
To capture similarities among classes that are not hierarchically related
Specifically, as an example, I've also used protocols to implement a provider model where I abstract out complete parts of the system. For example, to decouple my code a bit more, I could create a data provider. I could have a sqlite provider, a file provider etc... A protocol allows me to decouple those.
In that case, each of the providers do not inherit from each other but they all implement the same pattern with the internals being different. Think of a protocol as a pattern you conform to - or, a contract if you think of it like an interface.
Hope that helps.
You would use it anywhere you want to define a set of shared behavior without imposing a particular inheritance tree.
So my question is, why to use interfaces or abstract classes? Why are they useful, and for what?
Where can i use them intelligently?
Interfaces allow you to express what a type does without worrying about how it is done. Implementation can be changed at will without impacting clients.
Abstract classes are like interfaces except they allow you to provide sensible default behavior for methods where it exists.
Use and examples depend on the language. If you know Java, you can find examples of both interfaces and abstract classes throughout the API. The java.util collections have plenty of both.
They're useful when you want to specify a set of common methods and properties that all classes that implement/inherit from them have, exposed behaviors that all should provide.
Particularly about interfaces, a class can implement multiple interfaces, so this comes in handy when you're trying to model the fact that its instances must exhibit multiple types of behavior.
Also, as Wikipedia puts it, an interface is a type definition: anywhere an object can be passed as parameter in a function or method call, the type of the object to be exchanged can be defined in terms of an interface instead of a specific class, this allowing later to use the same function exchanging different object types: hence such code turns out to be more generic and reusable.
Can anyone explain the differences between Protocols and Categories in Objective-C? When do you use one over the other?
A protocol is the same thing as an interface in Java: it's essentially a contract that says, "Any class that implements this protocol will also implement these methods."
A category, on the other hand, just binds methods to a class. For example, in Cocoa, I can create a category for NSObject that will allow me to add methods to the NSObject class (and, of course, all subclasses), even though I don't really have access to NSObject.
To summarize: a protocol specifies what methods a class will implement; a category adds methods to an existing class.
The proper use of each, then, should be clear: Use protocols to declare a set of methods that a class must implement, and use categories to add methods to an existing class.
A protocol says, "here are some methods I'd like you to implement." A category says, "I'm extending the functionality of this class with these additional methods."
Now, I suspect your confusion stems from Apple's use of the phrase "informal protocol". Here's the key (and most confusing) point: an informal protocol is actually not a protocol at all. It's actually a category on NSObject. Cocoa uses informal protocols pervasively to provide interfaces for delegates. Since the #protocol syntax didn't allow optional methods until Objective-C 2.0, Apple implemented optional methods to do nothing (or return a dummy value) and required methods to throw an exception. There was no way to enforce this through the compiler.
Now, with Objective-C 2.0, the #protocol syntax supports the #optional keyword, marking some methods in a protocol as optional. Thus, your class conforms to a protocol so long as it implements all the methods marked as #required. The compiler can determine whether your class implements all the required methods, too, which is a huge time saver. The iPhone SDK exclusively uses the Objective-C 2.0 #protocol syntax, and I can't think of a good reason not to use it in any new development (except for Mac OS X Cocoa apps that need to run on earlier versions of Mac OS X).
Categories:
A category is a way of adding new methods to all instances of an existing class without modifying the class itself.
You use a category when you want to add functionality to an existing class without deriving from that class or re-writing the original class.
Let's say you are using NSView objects in cocoa, and you find yourself wishing that all instances of NSView were able to perform some action. Obviously, you can't rewrite the NSView class, and even if you derive from it, not all of the NSView objects in your program will be of your derived type. The solution is to create a category on NSView, which you then use in your program. As long as you #import the header file containing your category declaration, it will appear as though every NSView object responds to the methods you defined in the catagory source file.
Protocols:
A protocol is a collection of methods that any class can choose to implement.
You use a protocol when you want to provide a guarantee that a certain class will respond to a specific set of methods. When a class adopts a protocol, it promises to implement all of the methods declared in the protocol header. This means that any other classes which use that class can be certain that those methods will be implemented, without needing to know anyting else about the class.
This can be useful when creating a family of similar classes that all need to communicate with a common "controller" class. The communication between the controller class and the controlled classes can all be packaged into a single protocol.
Side note: the objective-c language does not support multiple inheritance (a class can only derive from one superclass), but much of the same functionality can be provided by protocols because a class can conform to several different protocols.
To my understanding Protocols are a bit like Java's Interfaces. Protocols declare methods , but the implementation is up to each class. Categories seems to be something like Ruby's mixins. With Categories you can add methods to existing classes. Even built-in classes.
A protocol allows you to declare a list of methods which are not confined to any particular class or categories. The methods declared in the protocol can be adopted any class/categories. A class or category which adopts a protocol must implements all the required methods declared in the protocol.
A category allows you to add additional methods to an existing class but they do not allow additional instance variables. The methods the category adds become part of the class type.
Protocols are contracts to implement the specified methods. Any object that conforms to a protocol agrees to provide implementations for those methods. A good use of a protocol would be to define a set of callback methods for a delegate (where the delegate must respond to all methods).
Categories provide the ability to extend a current object by adding methods to it (class or instance methods). A good use for a category would be extending the NSString class to add functionality that wasn't there before, such as adding a method to create a new string that converts the receiver into 1337 5P34K.
NSString *test = #"Leet speak";
NSString *leet = [test stringByConvertingToLeet];
Definitions from S.G.Kochan's "Programming in Objective-C":
Categories:
A category provides an easy way for you to modularize the definition of a class into groups or categories of related methods. It also gives you an easy way to extend an existing class definition without even having access to the original source code for the class and without having to create a subclass.
Protocols:
A protocol is a list of methods that is shared among classes. The methods listed in the protocol do not have corresponding implementations; they’re meant to be implemented by someone else (like you!). A protocol provides a way to define a set of methods that are somehow related with a specified name. The methods are typically documented so that you know how they are to perform and so that you can implement them in your own class definitions, if desired.
A protocol list a set of methods, some of which you can optionally implement, and others that you are required to implement. If you decide to implement all of the required methods for a particular protocol, you are said to conform to or adopt that protocol. You are allowed to define a protocol where all methods are optional, or one where all are required.