Object prototypes

Prototypes are the mechanism by which JavaScript objects inherit features from one another, and they work differently than inheritance mechanisms in classical object-oriented programming languages. In this article we explore that difference, explain how prototype chains work, and look at how the prototype property can be used to add methods to existing constructors.

A prototype-based language?

JavaScript is often described as a prototype-based language — each object has a prototype object, which acts as a template object that it inherits methods and properties from. An object's prototype object may also have a prototype object, which it inherits methods and properties from, and so on. This is often referred to as a prototype chain, and explains why different objects have properties and methods defined on other objects available to them.

Well, to be exact, the properties and methods are defined on the prototype property on the Objects' constructor functions, not the object instances themselves.

In classic OOP, classes are defined, then when object instances are created all the properties and methods defined on the class are copied over to the instance. In JavaScript, they are not copied over — instead, a link is made between the object instance and its prototype (its __proto__ property, which is derived from the prototype property on the constructor), and the properties and methods are found by walking up the chain of prototypes.

Let's look at an example to make this a bit clearer.

Understanding prototype objects

Let's go back to the example in which we finished writing our Person() constructor — load the example in your browser. If you don't still have it from working through the last article, use our oojs-class-further-exercises.html example (see also the source code).

In this example, we have defined a constructor function, like so:

function Person(first, last, age, gender, interests) {
  // property and method definitions
}

We have then created an object instance like this:

var person1 = new Person('Bob', 'Smith', 32, 'male', ['music', 'skiing']);

If you type "person1." into your JavaScript console, you should see the browser try to auto-complete this with the member names available on this object:

In this list, you will see the members defined on person1's prototype object, which is the Person() constructor — name, age, gender, interests, bio, and greeting. You will however also see some other members — watch, valueOf, etc — these are defined on the Person() constructor's prototype object, which is Object. This demonstrates the prototype chain working.

So what happens if you call a method on person1, which is actually defined on Object? For example:

person1.valueOf()

This method simply returns the value of the object it is called on — try it and see! In this case, what happens is:

• The browser initially checks to see if the person1 object has a valueOf() method available on it.
• It doesn't, so the browser then checks to see if the person1 object's prototype object (Person) has a valueOf() method available on it.
• It doesn't either, so the browser then checks to see if the Person() constructor's prototype object (Object) has a valueOf() method available on it. It does, so it is called, and all is good!

The prototype property: Where inherited members are defined

So, where are the inherited properties and methods defined? If you look at the Object reference page, you'll see listed in the left hand side a large number of properties and methods — many more than the number of inherited members we saw available on the person1 object in the above screenshot. Some are inherited, and some aren't — why is this?

The answer is that the inherited ones are the ones defined on the prototype property (you could call it a sub-namespace) — that is, the ones that begin with Object.prototype., and not the ones that begin with just Object. The prototype property's value is an object, which is basically a bucket for storing properties and methods that we want to be inherited by objects further down the prototype chain.

So Object.prototype.watch(), Object.prototype.valueOf(), etc., are available to any object types that inherit from Object.prototype, including new object instances created from the constructor.

Object.is(), Object.keys(), and other members not defined inside the prototype bucket are not inherited by object instances or object types that inherit from Object.prototype. They are methods/properties available just on the Object() constructor itself.

1. You can check out existing prototype properties for yourself — go back to our previous example and try entering the following into the JavaScript console:

   Person.prototype

2. The output won't show you very much — after all, we haven't defined anything on our custom constructor's prototype! By default, a constructor's prototype always starts empty. Now try the following:

   Object.prototype

You'll see a large number of methods defined on Object's prototype property, which are then available on objects that inherit from Object, as shown earlier.

You'll see other examples of prototype chain inheritance all over JavaScript — try looking for the methods and properties defined on the prototype of the String, Date, Number, and Array global objects, for example. These all have a number of members defined on their prototype, which is why for example when you create a string, like this:

var myString = 'This is my string.';

myString immediately has a number of useful methods available on it, like split(), indexOf(), replace(), etc.

Revisiting create()

Earlier on we showed how the Object.create() method can be used to create a new object instance.

1. For example, try this in your previous example's JavaScript console:

   var person2 = Object.create(person1);

2. What create() actually does is to create a new object from a specified prototype object. Here, person2 is being created using person1 as a prototype object. You can check this by entering the following in the console:

   person2.__proto__

This will return the person1 object.

The constructor property

Every constructor function has a prototype property whose value is an object containing a constructor property. This constructor property points to the original constructor function. As you will see in the next section that properties defined on the Person.prototype property (or in general on a constructor function's prototype property, which is an object, as mentioned in the above section) become available to all the instance objects created using the Person() constructor. Hence, the constructor property is also available to both person1 and person2 objects.

1. For example, try these commands in the console:

   person1.constructor
   person2.constructor

   These should both return the Person() constructor, as it contains the original definition of these instances.

   A clever trick is that you can put parentheses onto the end of the constructor property (containing any required parameters) to create another object instance from that constructor. The constructor is a function after all, so can be invoked using parentheses; you just need to include the new keyword to specify that you want to use the function as a constructor.

2. Try this in the console:

   var person3 = new person1.constructor('Karen', 'Stephenson', 26, 'female', ['playing drums', 'mountain climbing']);

3. Now try accessing your new object's features, for example:

   person3.name.first
   person3.age
   person3.bio()

This works well. You won't need to use it often, but it can be really useful when you want to create a new instance and don't have a reference to the original constructor easily available for some reason.

The constructor property has other uses besides. For example, if you have an object instance and you want to return the name of the constructor it is an instance of, you can use the following:

instanceName.constructor.name

Try this, for example:

person1.constructor.name

Modifying prototypes

Let's have a look at an example of modifying the prototype property of a constructor function (methods added to the prototype are then available on all object instances created from the constructor).

1. Go back to our oojs-class-further-exercises.html example and make a local copy of the source code. Below the existing JavaScript, add the following code, which adds a new method to the constructor's prototype property:

   Person.prototype.farewell = function() {
     alert(this.name.first + ' has left the building. Bye for now!');
   };

2. Save the code and load the page in the browser, and try entering the following into the text input:

   person1.farewell();

You should get an alert message displayed, featuring the person's name as defined inside the constructor. This is really useful, but what is even more useful is that the whole inheritance chain has updated dynamically, automatically making this new method available on all object instances derived from the constructor.

Think about this for a moment. In our code we define the constructor, then we create an instance object from the constructor, then we add a new method to the constructor's prototype:

function Person(first, last, age, gender, interests) {
  // property and method definitions
}
var person1 = new Person('Tammi', 'Smith', 32, 'neutral', ['music', 'skiing', 'kickboxing']);
Person.prototype.farewell = function() {
  alert(this.name.first + ' has left the building. Bye for now!');
};

But the farewell() method is still available on the person1 object instance — its available functionality has been automatically updated. This proves what we said earlier about the prototype chain, and the browser looking upwards in the chain to find methods that aren't defined on the object instance itself rather than those methods being copied to the instance. This provides a very powerful, extensible system of functionality.

You will rarely see properties defined on the prototype property, because they are not very flexible when defined like this. For example you could add a property like so:

Person.prototype.fullName = 'Bob Smith';

But this isn't very flexible, as the person might not be called that. It'd be much better to do this, to build the fullName out of name.first and name.last:

Person.prototype.fullName = this.name.first + ' ' + this.name.last;

This however doesn't work, as this will be referencing the global scope in this case, not the function scope. Calling this property would return undefined undefined. This worked fine on the method we defined earlier in the prototype because it is sitting inside a function scope, which will be transferred successfully to the object instance scope. So you might define constant properties on the prototype (i.e. ones that never need to change), but generally it works better to define properties inside the constructor.

In fact, a fairly common pattern for more object definitions is to define the properties inside the constructor, and the methods on the prototype. This makes the code easier to read, as the constructor only contains the property definitions, and the methods are split off into separate blocks. For example:

// Constructor with property definitions
function Test(a, b, c, d) {
  // property definitions
};
// First method definition
Test.prototype.x = function() { ... }
// Second method definition
Test.prototype.y = function() { ... }
// etc.

This pattern can be seen in action in Piotr Zalewa's school plan app example.

Summary

This article has covered JavaScript object prototypes, including how prototype object chains allow objects to inherit features from one another, the prototype property and how it can be used to add methods to constructors, and other related topics.

In the next article we'll look at how you can implement inheritance of functionality between two of your own custom objects.