Class (computer programming)

In object-oriented programming, a class is a construct that is used to define a distinct type. The class is instantiated into instances of itself – referred to as class instances, class objects, instance objects or simply objects. A class defines constituent members that enable its instances to have state and behavior. Data field members (member variables or instance variables) enable a class instance to maintain state. Other kinds of members, especially methods, enable the behavior of class instances. Classes therefore define the type of their instances.

A class usually represents a noun, such as a person, place or thing, or something nominalized. For example, a "Banana" class would represent the properties and functionality of bananas in general. A single, particular banana would be an instance of the "Banana" class, an object of the type "Banana".

Design and implementation

Classes are composed from structural and behavioral constituents. Programming languages that include classes as a programming construct offer support for various class-related features, and the syntax required to use these features varies greatly from one programming language to another.

Structure

UML notation for classes

A class contains data field descriptions (or properties, fields, data members, or attributes). These are usually field types and names that will be associated with state variables at program run time; these state variables either belong to the class or specific instances of the class. In most languages, the structure defined by the class determines the layout of the memory used by its instances. Other implementations are possible: for example, objects in Python use associative key-value containers.

Some programming languages support specification of invariants as part of the definition of the class, and enforce them through the type system. Encapsulation of state is necessary for being able to enforce the invariants of the class.

Behavior

Main article: Method (computer programming)

The behavior of class or its instances is defined using methods. Methods are subroutines with the ability to operate on objects or classes. These operations may alter the state of an object or simply provide ways of accessing it. Many kinds of methods exist, but support for them varies across languages. Some types of methods are created and called by programmer code, while other special methods—such as constructors, destructors, and conversion operators—are created and called by compiler-generated code. A language may also allow the programmer to define and call these special methods.

The concept of class interface

Main article: Interface (computing)

Every class implements (or realizes) an interface by providing structure and behavior. Structure consists of data and state, and behavior consists of code that specifies how methods are implemented. There is a distinction between the definition of an interface and the implementation of that interface; however, this line is blurred in many programming languages because class declarations both define and implement an interface. Some languages, however, provide features that separate interface and implementation. For example, an abstract class can define an interface without providing implementation.

Languages that support class inheritance also allow classes to inherit interfaces from the classes that they are derived from. In languages that support access specifiers, the interface of a class is considered to be the set of public members of the class, including both methods and attributes (via implicit getter and setter methods); any private members or internal data structures are not intended to be depended on by external code and thus are not part of the interface.

Object-oriented programming methodology dictates that the operations of any interface of a class are to be independent of each other. It results in a layered design where clients of an interface use the methods declared in the interface. An interface places no requirements for clients to invoke the operations of one interface in any particular order. This approach has the benefit that client code can assume that the operations of an interface are available for use whenever the client has access to the object.

Example

The buttons on the front of your television set are the interface between you and the electrical wiring on the other side of its plastic casing. You press the "power" button to toggle the television on and off. In this example, your particular television is the instance, each method is represented by a button, and all the buttons together comprise the interface. (Other television sets that are the same model as yours would have the same interface.) In its most common form, an interface is a specification of a group of related methods without any associated implementation of the methods.

A television set also has a myriad of attributes, such as size and whether it supports color, which together comprise its structure. A class represents the full description of a television, including its attributes (structure) and buttons (interface).

Getting the total number of televisions manufactured could be a static method of the television class. This method is clearly associated with the class, yet is outside the domain of each individual instance of the class. Another example would be a static method that finds a particular instance out of the set of all television objects.

Member accessibility

Further information: Information hiding

Many languages support the concept of information hiding and encapsulation, typically with access specifiers for class members. Access specifiers control access to class members. Some access specifiers may also control how classes inherit such constraints. Their primary purpose is to separate the interface of a class from its implementation.

The following is a common set of access specifiers:

• private (or class-private) restricts the access to the class itself. Only methods that are part of the same class can access private members.
• protected (or class-protected) allows the class itself and all its subclasses to access the member.
• public means that any code can access the member by its name.

Although many object-oriented languages support the above access specifiers, their semantics may differ.

Object-oriented design uses the access specifiers in conjunction with careful design of public method implementations to enforce class invariants—constraints on the state of the objects. A common usage of access specifiers is to separate the internal data of a class from its interface: the internal structure is made private, while public accessor methods can be used to inspect or alter such private data.

Access specifiers do not necessarily control visibility, in that even private members may be visible to client external code. In some languages, an inaccessible but visible member may be referred to at run-time (for example, by a pointer returned from a member function), but an attempt to use it by referring to the name of the member from client code will be prevented by the type checker.

The various object-oriented programming languages enforce member accessibility and visibility to various degrees, and depending on the language's type system and compilation policies, enforced at either compile-time or run-time. For example, the Java language does not allow client code that accesses the private data of a class to compile. In the C++ language, private methods are visible, but not accessible in the interface; however, they may be made invisible by explicitly declaring fully abstract classes that represent the interfaces of the class.

Some languages feature other accessibility schemes:

• Instance vs. class accessibility: Ruby supports instance-private and instance-protected access specifiers in lieu of class-private and class-protected, respectively. They differ in that they restrict access based on the instance itself, rather than the instance's class.

• Friend: C++ supports a mechanism where a function explicitly declared as a friend function of the class may access the members designated as private or protected.

• Path-based: Java supports restricting access to a member within a Java package, which is roughly the path of a file.

Inter-class relationships

In addition to the design of standalone classes, programming languages may support more advanced class design based upon relationships between classes. The inter-class relationship design capabilities commonly provided are compositional and hierarchical.

Compositional

Classes can be composed of other classes, thereby establishing a compositional relationship between the enclosing class and its embedded classes. Compositional relationship between classes is also commonly known as a has-a relationship. For example, a class "Car" could be composed of and contain a class "Engine". Therefore, a Car has an Engine. One aspect of composition is containment, which is the enclosure of component instances by the instance that has them. If an enclosing object contains component instances by value, the components and their enclosing object have a similar lifetime. If the components are contained by reference, they may not have a similar lifetime. For example, in Objective-C 2.0:

@interface Car : NSObject

@property NSString *name;
@property Engine *engine
@property NSArray *tyres;

@end

This Car class has an instance of NSString (string object), Engine, and NSArray (array object).

Hierarchical

Classes can be derived from one or more existing classes, thereby establishing a hierarchical relationship between the derived-from classes (base classes, parent classes or superclasses) and the derived class (child class or subclass) . The relationship of the derived class to the derived-from classes is commonly known as an is-a relationship. For example, a class 'Button' could be derived from a class 'Control'. Therefore, a Button is a Control. Structural and behavioral members of the parent classes are inherited by the child class. Derived classes can define additional structural members (data fields) and/or behavioral members (methods) in addition to those that they inherit and are therefore specializations of their superclasses. Also, derived classes can override inherited methods if the language allows.

Not all languages support multiple inheritance. For example, Java allows a class to implement multiple interfaces, but only inherit from one class. If multiple inheritance is allowed, the hierarchy is a directed acyclic graph (or DAG for short), otherwise it is a tree. The hierarchy has classes as nodes and inheritance relationships as links. Classes in the same level are more likely to be associated than classes in different levels. The levels of this hierarchy are called layers or levels of abstraction.

Example (Simplified Objective-C 2.0 code, from iPhone SDK):

@interface UIResponder : NSObject //...
@interface UIView : UIResponder //...
@interface UIScrollView : UIView //...
@interface UITableView : UIScrollView //...

In this example, a UITableView is a UIScrollView is a UIView is a UIResponder is an NSObject.

Definitions of subclass

Main articles: Inheritance (object-oriented programming), Superclass (computer science), and Subclass (computer science)

Conceptually, a superclass should be considered as a common part of its subclasses. This factoring of commonality is one mechanism for providing reuse. Thus, extending a superclass by modifying the existing class is also likely to narrow its applicability in various situations. In object-oriented design, careful balance between applicability and functionality of superclasses should be considered. Subclassing is different from subtyping in that subtyping deals with common behavior whereas subclassing is concerned with common structure.

There are two different points of view as to whether subclasses of the same class are required to be disjoint. Sometimes, subclasses of a particular class are considered to be completely disjoint. That is, every instance of a class has exactly one most-derived class, which is a subclass of every class that the instance has. This view does not allow dynamic change of object's class, as objects are assumed to be created with a fixed most-derived class. The basis for not allowing changes to object's class is that the class is a compile-time type, which does not usually change at runtime, and polymorphism is utilized for any dynamic change to the object's behavior, so this ability is not necessary. Design that does not need to perform changes to object's type will be more robust and easy-to-use from the point of view of the users of the class.

From another point of view, subclasses are not required to be disjoint. Then there is no concept of a most-derived class, and all types in the inheritance hierarchy that are types of the instance are considered to be equally types of the instance. This view is based on a dynamic classification of objects, such that an object may change its class at runtime. Then object's class is considered to be its current structure, but changes to it are allowed. The basis for allowing changes to object's class is a perceived inconvenience caused by replacing an instance with another instance of a different type, since this would require change of all references to the original instance to be changed to refer to the new instance. When changing the object's class, references to the existing instances do not need to be replaced with references to new instances when the class of the object changes. However, this ability is not readily available in all programming languages. This analysis depends on the proposition that dynamic changes to object structure are common. This may or may not be the case in practice.

Orthogonality of the class concept and inheritance

Although class-based languages are commonly assumed to support inheritance, inheritance is not an intrinsic aspect of the concept of classes. Some languages, often referred to as "object-based languages", support classes yet do not support inheritance. Examples of object-based languages include earlier versions of Visual Basic.

Within object-oriented analysis

Main article: Association (object-oriented programming)

In object-oriented analysis and in UML, an association between two classes represents a collaboration between the classes or their corresponding instances. Associations have direction; for example, a bi-directional association between two classes indicates that both of the classes are aware of their relationship. Associations may be labeled according to their name or purpose.

An association role is given end of an association and describes the role of the corresponding class. For example, a "subscriber" role describes the way instances of the class "Person" participate in a "subscribes-to" association with the class "Magazine". Also, a "Magazine" has the "subscribed magazine" role in the same association. Association role multiplicity describes how many instances correspond to each instance of the other class of the association. Common multiplicities are "0..1", "1..1", "1..*" and "0..*", where the "*" specifies any number of instances.

Taxonomy of classes

There are many categories of classes; however, these categories do not necessarily divide classes into distinct partitions.

Abstract and Concrete

Main article: Abstract type

In a language that supports inheritance, an abstract class, or abstract base class (ABC), is a class that cannot be instantiated because it is either labeled as abstract or it simply specifies abstract methods (or virtual methods). Abstract classes specify virtual methods via signatures that are to be implemented by direct or indirect descendents of the abstract class. Before a class derived from an abstract class can be instantiated, all abstract methods of its parent classes must be implemented by some class in the derivation chain.

Most object oriented programming languages allow the programmer to specify which classes are considered abstract and will not allow these to be instantiated. For example, in Java and PHP, the keyword abstract is used. In C++, an abstract class is a class having at least one abstract method given by the appropriate syntax in that language (a pure virtual function in C++ parlance).

A class consisting of only virtual methods is called a Pure Abstract Base Class (or Pure ABC) in C++ and is also known as an interface by users of the language. Other languages, notably Java and C#, support a variant of abstract classes called an interface via a keyword in the language. In these languages, multiple inheritance is not allowed, but a class can implement multiple interfaces. Such a class can only contain abstract publicly accessible methods.

A concrete class is a class that can be instantiated, as opposed to abstract classes, which cannot.

Local and inner

In some languages, classes can be declared in scopes other than the global scope. There are various types of such classes.

An Inner class is a class defined within another class. The relationship between an inner class and its containing class can also be treated as another type of class association. An inner class is typically neither associated with instances of the enclosing class nor instantiated along with its enclosing class. Depending on language, it may or may not be possible to refer to the class from outside the enclosing class. A related concept is inner types, also known as inner data type or nested type, which is a generalization of the concept of inner classes. C++ is an example of a language that supports both inner classes and inner types (via typedef declarations).

Another type is a local class, which is a class defined within a procedure or function. This limits references to the class name to within the scope where the class is declared. Depending on the semantic rules of the language, there may be additional restrictions on local classes compared non-local ones. One common restriction is to disallow local class methods to access local variables of the enclosing function. For example, in C++, a local class may refer to static variables declared within its enclosing function, but may not access the function's automatic variables.

Metaclasses

Main article: Metaclass

Metaclasses are classes whose instances are classes. A metaclass describes a common structure of a collection of classes and can implement a design pattern or describe particular kinds of classes. Metaclasses are often used to describe frameworks.

In some languages, such as Python, Ruby or Smalltalk, a class is also an object; thus each class is an instance of a unique metaclass which is built into the language. For example, in Objective-C, each object and class is an instance of NSObject. The Common Lisp Object System (CLOS) provides metaobject protocols (MOPs) to implement those classes and metaclasses.

Non-subclassable

Non-subclassable classes allow programmers to design classes and hierarchies of classes which at some level in the hierarchy, further derivation is prohibited. (A stand-alone class may be also designated as non-subclassable, preventing the formation of any hierarchy). Contrast this to abstract classes, which imply, encourage, and require derivation in order to be used at all. A non-subclassable class is implicitly concrete.

A non-subclassable class is created by declaring the class as sealed in C# or as final in Java. For example, Java's String class is designated as final.

Non-subclassable classes may allow a compiler (in compiled languages) to perform optimizations that are not available for subclassable classes.

Partial

In languages supporting the feature, a partial class is a class whose definition may be split into multiple pieces, within a single source-code file or across multiple files. The pieces are merged at compile-time, making compiler output the same as for a non-partial class.

The primary motivation for introduction of partial classes is to facilitate the implementation of code generators, such as visual designers. It is otherwise a challenge or compromise to develop code generators that can manage the generated code when it is interleaved within developer-written code. Using partial classes, a code generator can process a separate file or coarse-grained partial class within a file, and is thus alleviated from intricately interjecting generated code via extensive parsing, increasing compiler efficiency and eliminating the potential risk of corrupting developer code. In a simple implementation of partial classes, the compiler can perform a phase of precompilation where it "unifies" all the parts of a partial class. Then, compilation can proceed as usual.

Other benefits and effects of the partial class feature include:

• Enables separation of a class's interface and implementation code in a unique way.
• Eases navigation through large classes within an editor.
• Enables separation of concerns, in a way similar to aspect-oriented programming but without using any extra tools.
• Enables multiple developers to work on a single class concurrently without the need to merge individual code into one file at a later time.

Partial classes have existed in Smalltalk under the name of Class Extensions for considerable time. With the arrival of the .NET framework 2, Microsoft introduced partial classes, supported in both C# 2.0 and Visual Basic 2005. WinRT also supports partial classes.

Example in VB.NET

This simple example, written in Visual Basic .NET, shows how parts of the same class are defined in two different files.

file1.vb

Partial Class MyClass
    Private _name As String
End Class

file2.vb

Partial Class MyClass
    Public Readonly Property Name() As String
         Get
             Return _name
         End Get
    End Property
End Class

When compiled, the result is the same as if the two files were written as one, like this:

Class MyClass
    Private _name As String
    Public Readonly Property Name() As String
         Get cu
             Return _name
         End Get
    End Property
End Class

Example in Objective-C

In Objective-C, partial classes, aka categories may even spread over multiple libraries and executables, like this example:

In Foundation, header file NSData.h:

@interface NSData : NSObject

- (id)initWithContentsOfURL:(NSURL *)URL;
//...

@end

In user-supplied library, a separate binary from Foundation framework, header file NSData+base64.h:

#import <Foundation/Foundation.h>

@interface NSData (base64)

- (NSString *)base64String;
- (id)initWithBase64String:(NSString *)base64String;

@end

And in an app, yet another separate binary file, source code file main.m:

#import <Foundation/Foundation.h>
#import "NSData+base64.h"

int main(int argc, char *argv[])
{
    if (argc < 2)
        return EXIT_FAILURE;
    NSString *sourceURLString = [NSString stringWithCString:argv[1]];
    NSData *data = [[NSData alloc] initWithContentsOfURL:[NSURL URLWithString:sourceURLString]];
    NSLog(@"%@", [data base64String]);
    return EXIT_SUCCESS;
}

The dispatcher will find both methods called over the NSData instance and invoke both of them correctly.

Uninstantiable

Uninstantiable classes allow programmers to group together per-class fields and methods that are accessible at runtime without an instance of the class. Indeed, instantiation is prohibited for this kind of class.

For example, in C#, a class marked "static" can not be instantiated, can only have static members (fields, methods, other), may not have instance constructors, and is sealed.

Unnamed

An unnamed class or anonymous class is a class which is not bound to a name or identifier upon definition. This is analogous to named versus unnamed functions.

Benefits

Computer programs usually model aspects of some real or abstract world (the Domain). Because each class models a concept, classes provide a more natural way to create such models. Each class in the model represents a noun in the domain, and the methods of the class represent verbs that may apply to that noun (Verbs can also be modeled as classes, see Command Pattern). For example in a typical business system, various aspects of the business are modeled, using such classes as Customer, Product, Worker, Invoice, Job, etc. An Invoice may have methods like Create, Print or Send, a Job may be Performed or Canceled, etc. Once the system can model aspects of the business accurately, it can provide users of the system with useful information about those aspects. Classes allow a clear correspondence (mapping) between the model and the domain, making it easier to design, build, modify and understand these models. Classes provide some control over the often challenging complexity of such models.

Classes can accelerate development by reducing redundant program code, testing and bug fixing. If a class has been thoroughly tested and is known to be a 'solid work', it is usually true that using or extending the well-tested class will reduce the number of bugs - as compared to the use of freshly developed or ad hoc code - in the final output. In addition, efficient class reuse means that many bugs need to be fixed in only one place when problems are discovered.

Another reason for using classes is to simplify the relationships of interrelated data. Rather than writing code to repeatedly call a graphical user interface (GUI) window drawing subroutine on the terminal screen (as would be typical for structured programming), it is more intuitive. With classes, GUI items that are similar to windows (such as dialog boxes) can simply inherit most of their functionality and data structures from the window class. The programmer then need only add code to the dialog class that is unique to its operation. Indeed, GUIs are a very common and useful application of classes, and GUI programming is generally much easier with a good class framework.

Run-time representation

As a data type, a class is usually considered as a compile-time construct. A language may also support prototype or factory metaobjects that represent run-time information about classes, or even represent metadata that provides access to reflection facilities and ability to manipulate data structure formats at run-time. Many languages distinguish this kind of run-time type information about classes from a class on the basis that the information is not needed at run-time. Some dynamic languages do not make strict distinctions between run-time and compile-time constructs, and therefore may not distinguish between metaobjects and classes.

For example, if Human is a metaobject representing the class Person, then instances of class Person can be created by using the facilities of the Human metaobject.

See also

• Object-oriented programming
• Class-based programming
• Prototype-based programming
• Class diagram (UML)
• List of object-oriented programming languages

References

• Booch, Grady (1994). Objects and Design with Applications, Second Edition. Benjamin/Cummings. {{cite book}}: Invalid |ref=harv (help)

• Gamma; Helm; Johnson; Vlissides (1995). Design Patterns: Elements of Reusable Object-Oriented Software. Addison Wesley. {{cite book}}: Invalid |ref=harv (help)

Further reading

• Abadi; Cardelli: A Theory of Objects
• ISO/IEC 14882:2003 Programming Language C++, International standard
• Class Warfare: Classes vs. Prototypes, by Brian Foote
• Meyer, B.: "Object-oriented software construction", 2nd edition, Prentice Hall, 1997, ISBN 0-13-629155-4
• Rumbaugh et al.: "Object-oriented modeling and design", Prentice Hall, 1991, ISBN 0-13-630054-5

External links

• Dias, Tiago (2006). "Programming demo - .NET using Partial Types for better code". Hyper/Net. Youtube. {{cite web}}: Unknown parameter |month= ignored (help)