What is the difference between Type and Class?

问题

What makes a type different from class and vice versa?

(In the general language-agnostic sense)

回答1:

The following answer is from Gof book (Design Patterns)

An object's class defines how the object is implemented .The class defines object's internal state and the implementation of its operations.

In contrast, an object's type only refers to its interface - a set of requests to which it can respond.

An object can have many types, and objects of different classes can have the same type.

//example in c++
template<typename T> 
const T & max(T const & a,T const &b)
{
return a>b?a:b;  //> operator of the type is used for comparison
}

max function requires a type with operation > with its own type as one of it interface any class that satisfies the above requirement can be used to generate specific max function for that class.

回答2:

I always think of a 'type' as an umbrella term for 'classes' and 'primitives'.

int foo; // Type is int, class is nonexistent.

MyClass foo; // Type is MyClass, class is MyClass

回答3:

Inspired by Wikipedia...

In type theory terms;

• A type is an abstract interface.
  Types generally represent nouns, such as a person, place or thing, or something nominalized,

• A class represents an implementation of the type.
  It is a concrete data structure and collection of subroutines

  Different concrete classes can produce objects of the same abstract type (depending on type system).

  *For example, one might implement the type Stack with two classes: SmallStack (fast for small stacks, but scales poorly) and ScalableStack (scales well but high overhead for small stacks).*

  Similarly, a given class may have several different constructors.

The banana example.

• A Banana type would represent the properties and functionality of bananas in general.

• The ABCBanana and XYZBanana classes would represent ways of producing bananas.
  (Different banana suppliers in real life, or different data structures and functions to represent and draw bananas in a video game).

  The ABCBanana class could then produce particular bananas which are instances of the ABCBanana class, they would be objects of type Banana.

It is not rare the programmer provide a single and only implementation for a type. In this case the class name is often identical with the type name. But there is still a type (which could be extracted in an interface if required), and an implementation (which would implement the separate interface) which builds instances (objects) of the class.

回答4:

Type is the umbrella term for all the available object templates or concepts. A class is one such object template. So is the structure type, the Integer type, the Interface type etc. These are all types

If you want, you can look at it this way: A type is the parent concept. All the other concepts: Class, Interface, Structure, Integer etc inherit from this concept.i.e They are types

回答5:

Type contains description of the data (i.e. properties, operations, etc),

Class is a specific type - it is a template to create instances of objects.

Strictly speaking class is a special concept, it can be seen as a package containing subset of metadata describing some aspects of an object.

For example in C# you can find interfaces and classes. Both of them are types, but interface can only define some contract and can not be instantiated unlike classes.

Simply speaking class is a specialized type used to encapsulate properties and behavior of an object.

Wikipedia can give you a more complete answer:

• Definition of class
• Definition of data type

回答6:

To illustrate it the fastest way:

A Struct is a Type, but a Struct is not a Class.

As you can see, a Type is an "abstract" term for not only definitions of classes, but also structs and primitive data types like float, int, bool.

回答7:

Type is conceptually a superset of class. In the broader sense, a class is one form of type.

Closely related to classes are interfaces, which can bee seen as a very special kind of class - a purely abstract one. These too are types.

So "type" encompasses classes, interfaces and in most languages primitives too. Also platforms like the dot-net CLR have structure types too.

回答8:

My thoughts are pretty much in line with aku's answer.

I see classes as a template for building objects, while types are a way to classify those objects, and provide us with an interface to them.

Python also adds metaclasses, that are just a mechanism to build classes, in the same way as classes build objects (and well, classes and metaclasses are both objects).

This response to the same question in lamba the ultimate seems to me like a perfect explanation.

回答9:

Taken from the GoF citation from below:

An objects's class defines how the object is implemented .The class defines the object's internal state and the implementation of its operations.

In contrast, an objects's type only refers to its interface -the set of requests to which it can respond.

I want to provide an example using Java:

public interface IType {
}

public class A implements IType {
public A{};
}

public class B implements IType {
public B{};
}

Both classes A and B implement the interface and thus are of the type IType. Additionally in Java, both classes produce their own type (respectively to their class name). Thus the class A is of type A and IType and the class B is of type B and IType satisfying:

An object can have many types, and objects of different classes can have the same type.

The difference between subtypes and subclass probably helps to understand that issue as well:

https://www.cs.princeton.edu/courses/archive/fall98/cs441/mainus/node12.html

回答10:

To add another example of distinction: in C++ you have pointer and reference types which can refer to classes, but are not classes in and of themselves.

Bar b; // b is of type "class Bar"
Bar *b2 = &b; // b2 is of type "pointer to Class Bar"
Bar &b3 = b; // b3 is of type "reference to Class Bar"
Bar *b4[7]; // b4 is of type "7-element array of pointers to Class Bar"
Bar ***b5; //b5 is of type "pointer to a pointer to a pointer to Class Bar"

Note that only one class is involved, but a near infinite number of types can be used. In some languages, function are considered "first-class-objects" in which case, the type of a function is a class. In others, the type of a function is merely a pointer. Classes generally have the concepts of being able to hold data, as well as operations on that data.

回答11:

I think of a type as being the set of things you can do with a particular value. For instance, if you have an integer value, you can add it to other integers (or perform other arithmetic operations), or pass it to functions which accept an integer argument. If you have an object value, you can call methods on it that are defined by its class.

Because a class defines what you can do with objects of that class, a class defines a type. A class is more than that though, since it also provides a description of how the methods are implemented (something not implied by the type) and how the fields of the object are laid out.

Note also that an object value can only have one class, but it may have multiple types, since every superclass provides a subset of the functionality available in the object's class.

So although objects and types are closely related, they are really not the same thing.

回答12:

Type generally refers to the classification of primitive values - integers, strings, arrays, booleans, null, etc. Usually, you can't create any new types.

Class refers to the named set of properties and methods which an object is associated with when it is created. You can usually define as many new classes as you want, although some languages you have to create a new object and then attach methods to it.

This definition is mostly true, but some languages have attempted to combine types and classes in various ways, with various beneficial results.

回答13:

Types and classes are related but not identical. My take is that classes are used for implementation inheritance, whereas types are used for runtime substitution.

Here is a link explaining the substitution principle and why subclasses and subtypes are not always the same thing (in Java for example). The wikipedia page on covariance and contravariance has more information on this distinction.

回答14:

Interesting question. I think aku's answer is spot on. Take the java ArrayList class as an example

public class ArrayList<E> extends AbstractList<E>
    implements List<E>, RandomAccess, Cloneable, java.io.Serializable

An instance of the ArrayList class is said to be of type of every superclass it extends and every interface it implements. Therefore, an instance of the ArrayList class has a type ArrayList, RandomAccess, Cloneable, and so forth. In other words, values (or instances) belong to one or more types, classes define what these types are.

回答15:

Different classes may describe the same type.

Type consists of these parts:

1. Operations = syntax
2. Description of operations = semantics

Class consists of these parts:

1. Operations = syntax
2. Implementation (= various implementations describe same semantics)

Some notes:

• Interface (as in Java) is not type, because it does not describe semantics (describes only syntax)

• Subclass is not subtype, because subclass may change semantics defined in superclass, subtype cannot change supertype semantics (see Liskov Substitution Principle, e.g. this LSP example).

回答16:

Obviously, as there are languages with type system that are not OO programming languages, type must be a broader concept than class

Even in languages like Java, int is a (primitive) type, but not a class.

Hence: every class is a type, but not every type is a class.

回答17:

If we think to this question in C# context, we reach bellow answer.

C# type system is divided into following categories:

Value types:

• Simple types: like int, long, float, etc.
• Enum types
• Struct types
• Nullable types

Reference types:

• Class types
• Interface types
• Array types
• Delegate types

As you can see there are many types in C# which Class is only one of them. There is just one important note: C#’s type system is unified such that a value of any type can be treated as an object. Every type in C# directly or indirectly derives from the object class type, and object is the ultimate base class of all types. Values of reference types are treated as objects simply by viewing the values as type object. Values of value types are treated as objects by performing boxing and unboxing operations.

so as I see, type is an umbrella over many items which class is one of them.

Referece: CSahrp Language Specification doc, page 4

回答18:

In general language-agnostic sense - Class is an realization of the Type.

Often when this is the only realization of that type, you can use both terms to reference it in some context.

On the contrary, for example, in C# context - Class is just one of the many more implementations of a Type concept like primitives, structs, pointers etc.

回答19:

This was a good question for me, which made me think hard. I would dare to say that Class is a compiletime thingy and Type is a runtime thingy. I say this because you write classes not types. The compiler then creates types from classes, and the runtime use types to create instances of objects.

回答20:

Types in C, like Int Float, char etc define data that can be acted on with specific methods that can operate on them. It's no more complicated than that. Like for int I can add, subtract multiply and maybe divide. Those are my methods (or operations) for int. A Class is simply a definition of a new type. I first define what the data looks like. Maybe its a single bit. Maybe it's two words like a complex with a real and imaginary part. Or maybe its this complex thingy with 309734325 bytes representing the atomic makeup of a weird particle on Jupiter. I don't care. Just like an integer, I get to make up the operations I can do with this new data type. In the case of the integer I had add, subtract, etc. With this new data type I can define whatever operations I think make sense. They might be add subtract etc. but they may add other things. These are whatever methods I decide to add to my class.

The bottom line is that with a type in C, you have a definition of what the data is, ie; a byte, word, float, char etc. But any of these also implies what operations are legal and will produce reliable results.

A class is no different except it is up to you to define the interface and acceptable operations. The class defines these things and when you instantiate it in an Object it defines the behavior of the object just like a type definition defines the behavior of an integer when you operate on it.

Classes just give you the flexibility to define new types and everything about how they operate.

Once this is defined, every time I instantiate an object of class "thingy", it has the data structure I defined and the operations (methods) that I said you can do with it. The class "thingy" is clearly nothing more or less than a new type that C++ lets me define.

来源：https://stackoverflow.com/questions/468145/what-is-the-difference-between-type-and-class