Quick guide to somewhat advanced JavaScript

tour of some OO features

Hey, I didn't know you could do that

If you are a web developer and come from the same place I do, you have probably used quite a bit of Javascript in your web pages, mostly as UI glue.

Until recently, I knew that Javascript had more OO capabilities than I was employing, but I did not feel like I needed to use it. As the browsers started to support a more standardized featureset of Javascript and the DOM, it became viable to write more complex and functional code to run on the client. That helped giving birth to the AJAX phenomena.

As we all start to learn what it takes to write our cool, AJAXy applications, we begin to notice that the Javascript we used to know was really just the tip of the iceberg. We now see Javascript being used beyond simple UI chores like input validation and frivolous tasks. The client code now is far more advanced and layered, much like a real desktop application or a client-server thick client. We see class libraries, object models, hierarchies, patterns, and many other things we got used to seeing only in our server side code.

In many ways we can say that suddenly the bar was put much higher than before. It takes a heck lot more proficiency to write applications for the new Web and we need to improve our Javascript skills to get there. If you try to use many of the existing javascript libraries out there, like Prototype.js, Scriptaculous, moo.fx, Behaviour, YUI, etc you'll eventually find yourself reading the JS code. Maybe because you want to learn how they do it, or because you're curious, or more often because that's the only way to figure out how to use it, since documentation does not seem to be highly regarded with most of these libraries. Whatever the case may be, you'll face some kung-fu techniques that will be foreign and scary if you haven't seen anything like that before.

The purpose of this article is precisely explaining the types of constructs that many of us are not familiar with yet.

Related article

Prototype.js documentation

JSON

JavaScript Object Notation (JSON,) is one of the new buzzwords popping up around the AJAX theme. JSON, simply put, is a way of declaring an object in javascript. Let's see an example right away and note how simple it is.

var myPet = { color: 'black', leg_count: 4, communicate: function(repeatCount){ 
for(i=0;i<repeatCount;i++) alert('Woof!');} };
			

Let's just add little bit of formatting so it looks more like how we usually find out there:

var myPet = {
	color: 'black', 
	legCount: 4, 
	communicate: function(repeatCount){
		for(i=0;i<repeatCount;i++)
			alert('Woof!');
	} 
};
			

Here we created a reference to an object with two properties (color and legCount) and a method (communicate.) It's not hard to figure out that the object's properties and methods are defined as a comma delimited list. Each of the members is introduced by name, followed by a colon and then the definition. In the case of the properties it is easy, just the value of the property. The methods are created by assigning an anonymous function, which we will explain better down the line. After the object is created and assigned to the variable myPet, we can use it like this:

alert('my pet is ' + myPet.color);
alert('my pet has ' + myPet.legCount + ' legs');
//if you are a dog, bark three times:
myPet.communicate(3);
			

You'll see JSON used pretty much everywhere in JS these days, as arguments to functions, as return values, as server responses (in strings,) etc.

What do you mean? A function is an object too?

This might be unusual to developers that never thought about that, but in JS a function is also an object. You can pass a function around as an argument to another function just like you can pass a string, for example. This is extensively used and very handy.

Take a look at this example. We will pass functions to another function that will use them.

var myDog = {
	bark: function(){
		alert('Woof!');
	}
};

var myCat = {
	meow: function(){
		alert('I am a lazy cat. I will not meow for you.');
	}
};
 
function annoyThePet(petFunction){
	//let's see what the pet can do
	petFunction();
}

//annoy the dog:
annoyThePet(myDog.bark);
//annoy the cat:
annoyThePet(myCat.meow);
			

Note that we pass myDog.bark and myCat.meow without appending parenthesis "()" to them. If we did that we would not be passing the function, rather we would be calling the method and passing the return value, undefined in both cases here.

If you want to make my lazy cat start barking, you can easily do this:

myCat.meow = myDog.bark;
myCat.meow(); //alerts 'Woof!'
			

Arrays, items, and object members

The following two lines in JS do the same thing.

var a = new Array();
var b = [];
			

As I'm sure you already know, you can access individual items in an array by using the square brackets:

var a = ['first', 'second', 'third'];
var v1 = a[0];
var v2 = a[1];
var v3 = a[2];
			

But you are not limited to numeric indices. You can access any member of a JS object by using its name, in a string. The following example creates an empty object, and adds some members by name.

var obj = {}; //new, empty object
obj['member_1'] = 'this is the member value';
obj['flag_2'] = false;
obj['some_function'] = function(){ /* do something */};
			

The above code has identical effect as the following:

var obj = {
	member_1:'this is the member value',
	flag_2: false,
	some_function: function(){ /* do something */}
};
			

In many ways, the idea of objects and associative arrays (hashes) in JS are not distinguishable. The following two lines do the same thing too.

obj.some_function();
obj['some_function']();
			

Enough about objects, may I have a class now?

The great power of object oriented programming languages derive from the use of classes. I don't think I would have guessed how classes are defined in JS using only my previous experience with other languages. Judge for yourself.

//defining a new class called Pet
var Pet = function(petName, age){
	this.name = petName;
	this.age = age;
};

//let's create an object of the Pet class
var famousDog = new Pet('Santa\'s Little Helper', 15);
alert('This pet is called ' + famousDog.name);
			

Let's see how we add a method to our Pet class. We will be using the prototype property that all classes have. The prototype property is an object that contains all the members that any object of the class will have. Even the default JS classes, like String, Number, and Date have a prototype object that we can add methods and properties to and make any object of that class automatically gain this new member.

Pet.prototype.communicate = function(){ 
	alert('I do not know what I should say, but my name is ' + this.name);
};
			

That's when a library like prototype.js comes in handy. If we are using prototype.js, we can make our code look cleaner (at least in my opinion.)

var Pet = Class.create();
Pet.prototype = {
	//our 'constructor'
	initialize: function(petName, age){
		this.name = petName;
		this.age = age;
	},
	
	communicate: function(){
		alert('I do not know what I should say, but my name is ' + this.name);
	}
};	
			

Functions as arguments, an interesting pattern

If you have never worked with languages that support closures, like Ruby or C#2.0, you may find the following idiom too funky.

var myArray = ['first', 'second', 'third'];
myArray.each( function(item, index){
	alert('The item in the position #' + index + ' is:' + item);
});
			

Whoa! Let's explain what is going on here before you decide I've gone too far and navigate to a better article than this one.

First of all, in the above example we are using the prototype.js library, which adds the each function to the Array class. The each function accepts one argument that is a function object. This function, in turn, will be called once for each item in the array, passing two arguments when called, the item and the index for the current item. Let's call this function our iterator function. We could have also written the code like this.

function myIterator(item, index){
	alert('The item in the position #' + index + ' is:' + item);
}

var myArray = ['first', 'second', 'third'];
myArray.each( myIterator );
			

But then we would not be doing like all the cool kids in school, right? More seriously, though, this last format is simpler to understand but causes us to jump around in the code looking for the myIterator function. It's nice to have the logic of the iterator function right there in the same place it's called. Also, in this case, we will not need the iterator function anywhere else in our code, so we can transform it into an anonymous function without penalty.

Let's look at the original example again with some highlighting to hopefully make things clearer.

				
var myArray = ['first', 'second', 'third'];
myArray.each( function(item, index){
	alert('The item in the position #' + index + ' is:' + item);
} );
			

This is this but sometimes this is also that

One of the most common troubles we have with JS when we start writing our code is the use of the this keyword. It could be a real tripwire.

As we mentioned before, a function is also an object in JS, and sometimes we do not notice that we are passing a function around.

Take this code snippet as an example.

function buttonClicked(){
	alert('button ' + this.id + ' was clicked');
}

var myButton = document.getElementById('someButtonID');
var myButton2 = document.getElementById('someOtherButtonID');
myButton.onclick = buttonClicked;
myButton2.onclick = buttonClicked;
			

Because the buttonClicked function is defined outside any object we may tend to think the this keyword will contain a reference to the window or document object (assuming this code is in the middle of an HTML page viewed in a browser.)

But when we run this code we see that it works as intended and displays the id of the clicked button. What happened here is that we made the onclick method of each button contain the buttonClicked object reference, replacing whatever was there before. Now whenever the button is clicked, the browser will execute something similar to the following line.

myButton.onclick();
			

That isn't so confusing afterall, is it? But see what happens you start having other objects to deal with and you want to act on these object upon events like the button's click.

var myHelper = {
	
	formFields: [ ],
	
	emptyAllFields: function(){
		for(i=0; i<this.formFields.length; i++){
			var elementID = this.formFields[i];
			var field = document.getElementById(elementID);
			field.value = '';
		}
	}
};

//tell which form fields we want to work with
myHelper.formFields.push('txtName');
myHelper.formFields.push('txtEmail');
myHelper.formFields.push('txtAddress');

//clearing the text boxes:
myHelper.emptyAllFields();

var clearButton = document.getElementById('btnClear');
clearButton.onclick = myHelper.emptyAllFields;
			

So you think, nice, now I can click the Clear button on my page and those three text boxes will be emptied. Then you try clicking the button only to get a runtime error. The error will be related to (guess what?) the this keyword. The problem is that this.formFields is not defined if this contains a reference to the button, which is precisely what's happening. One quick solution would be to rewrite our last line of code.

clearButton.onclick = function(){ myHelper.emptyAllFields(); };
			

That way we create a brand new function that calls our helper method within the helper object's context.

---

Books I've read and recommend

Some books are just too good not to pass the word forward. I haven't read a whole lot of Javascript books, but the following books were very helpful to me. They are very well written and didn't cause me to lose interest before reading half of it. I'm comfortable recommending them if you're on the market for a new book.

This article is not finished yet. If you find broken examples, incorrect information, broken links, etc., please and I'll try to fix it as soon as possible.

今天不打cs 2007-03-09 16:20 发表评论