Part 1: Backbone.js Deconstructed

Outline

1 - Article Overview

Not a lot of in-depth information from a single voice has been written about Backbone itself. We have varying degrees of information about how Backbone can be implemented to create an application. We have lots of opinions about how it does or does not fit into the Model/View/Whatever world. We have more than enough "to-do" demo applications, which can bring a layer of value when learning about application design. However, what I think is lacking are details about Backbone itself and what the actual code is doing. The documentation offers some degree of this as well as the annotated source, but I think more information is needed. Or, at the very least, a different voice documenting Backbone's function and parts.

In Part 1 and Part 2 of Backbone.js Deconstructed, we are going to examine the function of the Backbone Library at the code level.

Specifically, in Part 1 we will identify some helpful (dare I say 'required'?) prerequisites that make learning Backbone a rewarding endeavor. Then we will cover some critical details about Single Page Applications (SPAs) that are needed before one can truly understand the purpose and parts of Backbone. After we have a mental foundation for SPAs, we will dive into the code by systematically breaking down the details around Backbone.Events, creating Backbone Objects, the Backbone.Router() and Backbone.History(), and lastly, Backbone utilities. In Part 2 of Backbone.js Deconstructed, we will systematically uncover the purpose and functionality provided by Backbone.Model, Backbone.View, and Backbone.Collection.

Now, for the part where I you tell you what you won't find in these two articles. You won't find in-depth information about MVC, MVVM, MVP, or MV-Whatever patterns. I purposefully have sidestepped any in-depth discussion about these patterns because I believe that Backbone is such a simple library that if you learn the function of the code, the implementation part will be obvious and you will remain free to adapt Backbone to your own domain.

You also won’t find an application being built in this article. This is commonly the method of teaching Backbone to newcomers and plenty of approaches in this fashion are available. Additionally, I am not writing this article to produce yet another tutorial that briefly talks about Models and Views.

In this article, we are breaking down the major parts of Backbone and examining the function of each part at the code level. Expect to get intimate with each part of Backbone.

2 - Backbone Prerequisites

2.1 - You Need Intermediate Knowledge of JavaScript and the DOM

Backbone abstracts a great deal of the repetitive tasks that can be accomplished with native DOM and HTML5 JavaScript APIs. Because of this, if you are not experienced with the DOM or with JavaScript, learning about a solution like Backbone might prove to be rather difficult, slow, and frustrating. Consider that Backbone itself reaches out for a third party solution (i.e. jQuery, Ender, or Zepto) for DOM manipulation and that is two layers of abstraction to get lost in if you do not already have a firm grip on native DOM scripting and modern JavaScript APIs (e.g. XHR). If you are not at least an intermediate JavaScript developer, you will mostly likely struggle with learning Backbone, as well as my words on the matter.

2.2 - You Need Experience with HTTP, XHR, and RESTful JSON APIs

Backbone comes with some helpers that make saving/syncing data to the backend trivial. Of course, a JSON API is assumed, so if you are not familiar with HTTP, XHR, or reading and writing data to the server by using a RESTful JSON API, you might not clearly understand what Backbone is doing for you, or its value in doing it so that you don't have to.

2.3 - Experience with a Functional Programming Library

When Backbone was authored, a supporting library was created along with it, called Underscore.js, which provides the functional programming utilities used by Backbone. This library is a separate tool, which can be used without Backbone, but Backbone has a hard dependency on Underscore.js. Having experience with a functional programming utility library like Underscore.js or Lo-Dash.js will help you not only understand the Backbone code better, it will also help you understand the functional methods that are inherited and usable on the Backbone Model and Collection objects. If you have not spent any time gleaning the programming goodness found in Underscore.js or Lo-Dash.js, I highly suggest you spent some time in the trenches of one of these libraries.

Notes:

2.4 - You Need a Loose Understanding of Models and Views

Backbone strives to alleviate the “tangled piles of jQuery selectors and callbacks, all trying frantically to keep data in sync between the HTML UI”, and in the right hands, it does just that. However, it is just as easy to create a tangled pile of Backbone Events, Models, Collections, and Views. One can always just jump in and thrash about until they figure out how to swim in the Backbone pool, but a general understanding of a model and view will aid a developer when it comes to learning and implementing Backbone. Unfortunately, gaining even a loose conceptual understanding of application architecture patterns and sub-design patterns is challenging. I would recommend, at the very least, a general understanding of a model and view without getting tied up in a flavor of architecture for grounding the concept of a model or view. If you are starting with zero knowledge about a model or a view, I am going to offer a short Backbone-biased definition for each. If you feel strongly that you already have a conceptual understanding of a model and view, skip this section so I don't taint your previous knowledge.

Model - A model contains the state (i.e. the JSON data or what Backbone calls attributes) for an application, as well as a basic set of functionality (i.e. methods like .set() or .get() and events like change) for managing and listening to changes in the state/data. Additionally, the model orchestrates the storing of this state/data in memory in the browser, as well as persisting the state/data to the database on the server (i.e. syncing). An ordered grouping of models is a collection. In Backbone, collections are wrapped with a basic set of functionality for working with groups of models. The classic illustration of a model would be the data for a to-do item (i.e. the to-do and its state of completion), and a list of to-dos would be considered a collection of to-do models. If you think in terms of a table, the table would be the collection (i.e. to-dos), the headers in the table would be the model (i.e. title, status), and the content in the table rows would be the data (i.e. get a dog, done).

todo's (i.e. collection of to-do models)
title (i.e. model attribute) status (i.e. model attribute)
get a dog (i.e. model attribute value) done (i.e. model attribute value)

View - A view contains the logic to render an interactive UI of the state/data contained in models and collections of models. Views allow for state/data to be decoupled from the DOM so that multiple views can make use of the same models and collections. Thus, if you change a model or a model contained in a collection, multiple views can listen for these changes and update accordingly. So, "the truth", or the state/data is not tied to the view; rather, it is decoupled from the view, or specifically, from the DOM/HTML.

3 - Application Architecture, Model View Whatever, and Backbone

3.1 - Architectural Assumptions Made by Backbone

Backbone was written for use in the context of developing a SPA (AKA Single Page Application). You might wonder what that is, exactly. In short, a SPA is a web architecture which loads a single web page (typically one index.html or app.html) through which the entire application can be delivered to the client by using custom events, custom url handling, and asynchronous HTTP retrieval and storing of resources, all while avoiding traditional browser page reloads. Basically, a SPA architecture loads one HTML page so that JavaScript can then be used to mimic/override/orchestrate the default nature of the browser that is used in traditional web application architectures. Let's examine the architectural assumptions of a Backbone SPA in more detail:

Single page applications which leverage RESTful JSON APIs are the sweet spot for Backbone. If you are building something else, or the data API is not set up for REST, you might still find value in the event, model, view, and collection logic provided by Backbone, but you should be aware that you are out of the Backbone sweet spot, given that many Backbone features assume a RESTful JSON API.

3.2 - Backbone Doesn’t (and You Shouldn’t) Focus on the Whatever part of Model-View-Whatever (i.e. MV*)

The study of MV-Whatever patterns is not without merit. However, the study of Backbone in the context of the MV-Whatever family might do more harm than good. Backbone is an incredibly small library, and the approach one could take to grok a Model-View-Whatever architecture can be accomplished by setting aside this topic altogether and simply studying what Backbone is doing at the code level. Instead of trying to wrap your head around the very opinionated concepts surrounding the 'Whatever' part of MV-Whatever (i.e. MV*), I recommend instead that you ground your understanding of Backbone in its function and the simple notion that its purpose is to provide aid when organizing application data (models) and UI (views) for SPAs. I believe that if you learn exactly what the JavaScript is doing, you will naturally start to grok the world of Backbone MV-Whatever.

3.3 - Backbone is a Library, Not a Framework

Backbone is a JavaScript library which provides an event system and a set of constructor functions for creating objects that can be used in a MV-Whatever manner to organize and sync data for a single page web application. The strongest framework-like-opinions built into Backbone are found in its assumption that a SPA architecture and RESTful JSON API are in use. Remember, a framework calls your code, while a library is code that you call. Backbone, in a nut shell, is a library of constructor functions that you leverage as a developer in order to add structure to your application. It is not a framework!

3.4 - Backbone Doesn't Provide Everything Needed to Develop a Single Page, Thick Client, Web Application

As previously mentioned, Backbone is not a framework. It is only a piece of the application puzzle. When building out a single page web application, you will likely require many of the developer tools listed below (the list does not include solutions for building a RESTful JSON API service):

If the list above looks like a foreign language to you, then the last thing I would recommend is learning Backbone via a third party Backbone framework that leverages these tools.

Before you use a Backbone framework, research each of the above tools. Use them and make sure you understand the problem they are solving. When this is done, check out some of the non-framework community extensions. Doing this will help you familiarize yourself with some of the gaps other developers have found when using Backbone.

When a proper understanding of Backbone itself has been achieved (along with its gaps), and you understand the problem each of the previously mentioned tools solves, you might find one of the following Backbone toolkits/frameworks useful:

However, inviting the additional layers of abstraction provided by these toolkits should be done with extreme caution. These solutions provide a great deal of boilerplate, opinions, and plenty of rope to hang yourself. I would only reach for one of these once you have a thorough understanding of Backbone and the underlining tools that these frameworks may use.

4 - Backbone Big Picture Stuff

4.1 - Backbone's Beginnings

Backbone was conjured up specifically to add structure to single page applications (to be exact, it was constructed during the creation of the DocumentCloud application) so as to avoid the repetitive DOM glue and mixing of data within the DOM that had historically been written to keep the UI in sync with the data. Backbone accomplishes this by providing a set of constructor functions which produce a model, collection, and view object, all for the purpose of organizing application data, logic, and presentation. These objects, once instantiated, have a very specific relationship to each other and are intended to keep your application modular, decoupled (communicating using an event system), and scalable.

4.2 - Backbone Provides Application Structure, as Well as Helpers for Building SPAs

In addition to the Backbone.Model(), Backbone.Collection(), and Backbone.View() constructor functions (i.e. the meat), Backbone additionally provides the following helpers for building SPAs and saving data with a RESTful JSON API:

These helper parts, just mentioned, should be understood before attempting to grok models, collections, and views. And, in fact, this is exactly what I have set out to accomplish in Part 1 of this article.

4.3 - Backbone Hard & Soft Dependencies

The only hard dependency identified by the Backbone documentation is the use of a functional programming utility library, either Underscore.js or Lo-Dash.js. Don’t get hung up here on which to choose, just choose one and use it, or Backbone literally won’t work. If you want to be told which to use, use Lo-Dash.js.

In addition to either Underscore.js or Lo-Dash.js, Backbone has soft dependencies on json2.js and your choice of jQuery, Zepto, or Ender (really anything that mimics the jQuery API will work). If the soft dependency on a DOM library is not included or window.JSON is not provided by the browser the Backbone.Router(), Backbone.View(), Backbone.History(), and Backbone.Sync() won't function properly. That is, if you don’t include the soft dependencies, Backbone only offers partial use of models and collections. I suppose that might be helpful to someone, but it's my opinion that 99 percent of the time, the soft dependencies will be included along with Backbone.js (depending upon what browsers you need to support).

An example HTML file that details the Backbone dependencies might look something like this:

<!DOCTYPE html>
<html lang="en">
<body>
<!-- functional programming utility library -->
<script src="http://cdnjs.cloudflare.com/ajax/libs/lodash.js/1.3.1/lodash.underscore.js"></script>
<!-- DOM & XHR utility -->
<script src="http://cdnjs.cloudflare.com/ajax/libs/jquery/2.0.2/jquery.min.js"></script>
<!-- needed for older browsers without window.JSON -->
<script src="http://cdnjs.cloudflare.com/ajax/libs/json2/20121008/json2.js"></script>
<!-- Backbone.js -->
<script src="http://cdnjs.cloudflare.com/ajax/libs/backbone.js/1.0.0/backbone-min.js"></script>
</body>
</html>

It is common, if not required, that you use a dependency manager when building out a web application (e.g. Require.js). The HTML example file above is not a recommendation. Rather, it is being shown to clearly communicate the Backbone hard and soft dependencies. My opinion is that a dependency manager should be used when building a single page web application (i.e. thick client application).

4.4 - Backbone Runs on the Server Too

Backbone is useful anywhere you need to structure an application. This includes server-side applications, as well. The developers behind Backbone have crafted the library so that it can be downloaded from NPM for Node.js server-side application development. However, its origins started in the client, and using it on the server will require some additional knowledge about running client-centric solutions from a JavaScript backend environment.

4.5 - Backbone is Just a Top-Level Namespace

The Backbone object that is exposed in the global scope is a namespace which contains Backbone code. Specifically, it contains the Model(), View(), Collection(), History(), and Router() constructor functions, as well as some supporting logic (i.e. Events and sync()) for these constructors. The code can be summarized like this:

(function(){

    var root = this;

    Backbone.VERSION = '1.0.0';

    Backbone.Events = {//events functions to mixed into any object};
    Backbone.Model = function() {//model constructor};
    Backbone.Collection = function() {//collection constructor};
    Backbone.Router = function() {//router constructor};
    Backbone.History = function() {//history constructor};
    Backbone.sync = function() {//sync function};
    Backbone.View = function() {//view constructor};

}).call(this);

5 - Backbone.Events

5.1 - Backbone.Events Overview

The Backbone.Events object contains methods that facilitate the observer design pattern and a variation of the observer pattern called pub/sub. These functions permit the various objects created by Backbone (really any JavaScript object) to communicate with each other in a decoupled manner. This is done by extending the prototypes (i.e. _.extend(Model.prototype, Backbone.Events);) of the following constructor functions with the functions contained in Backbone.Events:

If the observer design pattern is meaningless to you and the term pub/sub is a rather mussy concept, just think of the functions contained in Backbone.Events similar to functionality provided by the jQuery $('').on(), $('').off(), $('').trigger(), and $('').one() methods for attaching native or custom events to DOM objects with the added ability to listen to events as well.

The Backbone.Events object contains the following functions which create events, trigger events, and create listening relationships between objects based on events.

In addition to extending the prototype of each of the above constructor functions, the Backbone namespace object itself is extended (i.e. _.extend(Backbone, Backbone.Events);) with the properties found in Backbone.Events so that a global eventing system is available if needed. For example, below I created a generic sayHi event and an eavesdropper object to listen for any sayHi events that are triggered on the Backbone object.

Notes:

5.2 - Extending Any Object With Backbone.Events

As previously mentioned, the Backbone.Events object properties are mixed into the prototype chain of several of the constructor functions provided by Backbone, as well as the Backbone namespace object. Of course, nothing is stopping you from extending any object with the functions found in the Backbone.Events object. Backbone.Events basically is a standalone set of functions that can be mixed into any old JavaScript object. Below I demonstrate this by extending the A object and then verifying that the Backbone.Events methods are now properties of A.

5.3 - Binding Events to Objects using on()

The on() method will bind a callback function to an object that will be invoked whenever the event is triggered. In the code below I bind a whatsMyName event to the A object and then trigger that event.

Notes:

5.4 - Binding Multiple Events with on()

The on() accepts a string event parameter which can contain multiple events separated by a space. Below I add two events to the A object, which call the same function, and then I trigger this callback with the two different event names.

5.5 - Name Spacing Events with on()

It is possible to namespace events by using a colon in the event name followed by the namespace. Below I have added two say events with a unique namespace, which I then use to invoke these name spaced events.

5.6 - Triggering And Passing Values To A Callback Function Using trigger()

The trigger() method is used to invoke named (i.e. events) callback functions. Its use has been prevalent in this section. The second parameter passed to the trigger() method gives the option of passing values to the callback function. Below I trigger a callback function on A to which I pass several values.

Notes:

5.7 - Setting The this Context For A Callback Function Using on()

The third parameter that can be passed to on() allows the this context for the callback function to be set. In the code below I invoke a callback on A but tell the value of this in the callback to be B.

5.8 - Removing Events and Callback Functions with off()

The off() method can be used in the following 5 ways to remove individual events, callback functions, or all events from an object.

1 - Removing a named callback function:

2 - Removing all occurrences of an event:

3 - Removing all the occurrences of a named callback function from all events:

4 - Removing all events for a specific context:

5 - Removing all events and callbacks on an object:

Notes:

5.9 - Have an object listen for events on another object using listenTo()

The listenTo() method gives the ability for one object to listen for an event on another object, and when that event occurs provide a callback function that is invoked. In the code example below I tell object B to listen for the whosListeningToMe event to occur on object A, and when it does occur invoke the whosListeningToMe callback function.

Notes:

5.10 - Stop listening for events on another object using stopListening()

The stopListening() method can be used on objects that are listening to other objects in the following four ways:

1 - Stop All Listening

2 - Stop listening to a specific object

3 - Stop listening to a specific event on a specific object

4 - Stop Listening to a specific event and callback on a specific object

5.11 - Backbone Built-in Events

Backbone triggers the following internal events that have been built in to models, collections, views, and the router objects. Because Backbone triggers these events we can make use of these events by way of the on() and listenTo() Backbone event methods. More on that in a minute. For now, just make sure you review the description of individual events in the table below so you start to gain an understanding of when these events are triggered.

model, collection, view, router, and history object events

event type arguments passed to callback description
'all' event name this special event fires for any (i.e. not just built in events) triggered event, passing the event name as the first argument

collection object events

event type arguments passed to callback description
'add' model, collection, options when a model is added to a collection
'remove' model, collection, options when a model is removed from a collection
'reset' collection, options when the collection's entire contents have been replaced.
'sort' collection, options when the collection has been re-sorted

model object events

event type arguments passed to callback description
'change' model, options when a model's attributes have changed.
'change:[attribute]' model, value, options when a specific attribute has been updated.
'destroy' model, collection, options when a model is destroyed.
'error' model, xhr, options when a model's save call fails on the server.
'invalid' model, error, options when a model's validation fails on the client.

model or collection object events

event type arguments passed to callback description
'request' model, xhr, options when a model or collection has started a request to the server.
'sync' model, resp, options when a model (or collection) has been successfully synced with the 

router object events

event type arguments passed to callback description
'route:[name]' params Fired by the router when a specific route is matched.
'route' router, route, params Fired by history (or router) when any route has been matched.

The above events, except for the 'all' event, will be examined in more detail in the section that deals with the object that the event is triggered on. To gain an understanding about Backbone built in events we will examine the 'all' event now.

Built in events are triggered by Backbone objects. The 'all' event is triggered/broadcasted anytime any event is triggered on a Backbone model, collection, view, or router.

In the code below I instantiate a Backbone model then I bind (i.e on()) a callback to the 'all' event. Next, I set some data on the model. When this data is set, the model will broadcast two events, a 'change' event as well as a 'change:attribute' event (You can verify this in the table above). The 'all' event gets fired twice because it's broadcasted when events, in this case two 'change' events, are triggered.

The all event captures, as its first argument, the event that triggered the 'all' event. This allows us to log which event triggered the 'all' event.

Don't forget, the 'all' event is not unique to models it is triggered anytime any of the built in Backbone events are triggered. However, note that the 'change' events are specific to models. This example should lead the way for understanding broadly how all the built in events can be tapped into when developing Backbone applications.

5.12 - Backbone.Events Conclusion

The take away from this section is that the concepts used by Backbone objects to communicate with each other are generic patterns. By understanding the generic uses provided by Backbone.Events for any JavaScript object you will be better equipped to understand how to create your own events, as well as make use of the events that are internally triggered by Backbone Models, Collections, Views, and the Router.

6 - Constructing Backbone Objects

6.1 - Backbone Constructor Functions

When I claimed earlier that Backbone itself is simple, I did so because Backbone only provides 4 constructor functions that are typically instantiated (technically, they are extended or sub-classed first by a developer before instantiation).

These four constructors are:

Backbone.Model = function(attributes, {options}){};
Backbone.Collection = function([models], {options}){};
Backbone.Router = function({options}){}; //typically instantiated once 
Backbone.View = function({options}){};

Let's examine the default instance and prototype properties for each of the instances created from the above constructor functions by simply instantiating each with the new keyword (click on the results in the JSfiddles to view instance and prototype properties) passing no parameters.

Models:

Collection:

Router:

View:

The properties identified for each constructor function are what Backbone provides out of the box. However, it is more common that you will be extending (i.e. Backbone.Model.extend()) these constructors with your own domain specific instance and prototype properties and methods before you construct them.

Briefly, before we continue, let's examine the optional parameters for each constructor. More time will be spent on these parameters in the appropriate section but I think it's important to have a rough understanding of the optional parameters.

var myModel = new Backbone.Model(attributes, {options});
/*
attributes = {data:value, data:value} || [value,value]

options = {
 collection: {}, 
 url: '', 
 urlRoot: '', 
 parse: boolean
}
*/


var myCollection = new Backbone.Collection(model(s), {options});
/*
models = model || [model,model,model] || {model:data} || [{model:data},{model:data}];

options = {
 url: '', 
 model: {}, 
 comparator: function(){} || ''
}
*/


var myView = new Backbone.View({options});
/*
options = {
 model: {}, 
 events: {} or function(){return {}}
 collection: {}, 
 el: '' or function(){return ''}, 
 id: '', 
 className: '' or function(){return ''}, 
 tagName: '' or function(){return ''}, 
 attributes: {attribute:value,attribute:value}
}
*/


var myRouter = new Backbone.Router({options});
/*
options = {
 routes:{}
}
*/

/*should be aware of History options as well*/
Backbone.History.start({options});
/*
options = {
 pushState: boolean,
 root: '',
 hashChange: boolean,
 silent: boolean
}
*/

Notes:

6.2 - Using The Internal Backbone extend() Helper Function To Create Sub Classed Model(), Collection(), Router(), and View() Constructors.

The Backbone pattern for constructing a model, collection, view, or router routinely involves extending one of the default Backbone constructor functions (creating a sub class or sub constructor) with your own domain relevant properties and methods. This is accomplished with the extend() function that is a property of the constructor function. Don't confuse this function with the extend() provided by underscore.js.

The extend() provided by Backbone accomplishes, from a high level, the following:

For a more detailed break down of how the Backbone extend() method works, carefully read the comments in the code below which describes the logic contained in the extend() function for extending a fictitious View constructor (similar but not precisely Backbone.View). While this is not the exact code contained in Backbone.js its an isolated example of what is occuring at the code level when a Backbone.Model, Backbone.Collection, Backbone.View, and Backbone.Router constructor is sub classed (or what I like to call an extended sub constructor) .

Understanding extend(), and its role when using Backbone will take a lot of the mystery out of Backbone applications. I would suggest you examine its parts until you completely understand its purpose and function, but if you want to move past this subject quickly what you need to know about extend() boils down to one concept.

  1. extend(), creates or extends an object with new application specific object properties so that when the extended object is instantiated it will have these new instance properties and methods.

6.3 - Running Initialization Code When An Extended Constructor Is Instantiated

When extend()'ing Backbone.Model, Backbone.Collection, Backbone.View, and Backbone.Router an initialize function can be setup to run anytime an instance is created from the extended constructor. Basically, initialize is a callback, which is invoked any time an instance is instantiated. In the code below I provide an initialize callback for an extended Backbone.Model constructor and then create an instance, which in turn runs the initialized function (works the same on Backbone.Collection, Backbone.View, and Backbone.Router).

Make sure you note that any parameters passed to an extended constructor are also passed into the initialize callback.

Notes:

6.4 - Overriding The Constructor Being Extended

When using extend() if you pass a property called 'constructor' with a function as its value, this function will be used instead of the function the extend() is a property of. In the code below I am extending not the Model constructor, but the constructor function I am passing in.

In the code above myModel still inherits the Model() prototype properties, just not the instance properties (i.e. cid). If you override the constructor, but want to keep instance properties from the constructor you are calling extend() from, you need to run Backbone.[CONSTRUCTOR].apply(this, arguments); in the custom constructor*/

7 - Backbone.Router

7.1 - Backbone.Router() Overview

A Backbone route is a JavaScript string, which resembles what traditionally has been considered a URL pathname. This pathname-like string is the property name for a function (or reference to a named function) that is invoked when the url in the browser matches the string. For example, the url below contains the pathname "help" and in a single page Backbone application this would tell Backbone to invoke the function that is associated with the route called 'help'.

http://www.hostname.com/#help

Notice that the pathname "help" isn't actually a pathname, it is actually a hash. Backbone defaults to using hash paths because as of today its the most widely supported solution for creating unique URL strings without causing a browser to reload. When you no longer need to support ie 9 or below you can tell Backbone to use the newer HTML5 history.pushState which eliminates the need for hash paths. The history.pushState api updates the browser history and permits the modification of the URL pathname without causing the browser to reload. I will talk more about this later in this section. For now, let's look at an actual Backbone route.

In the code below I create an instance of the default Backbone.Router() and then define a "help" route using the route() method.

In the code example above when the hash changes to #help Backbone is aware of this change and takes everything after the # and matches that to a predefined string name (i.e. "help") which is associated with a callback function.

Routes are essentially pathname like strings in a URL that are monitored by Backbone for string matches. When a match is found a corresponding function is invoked.

Notes:

7.2 - Starting Routing Management with Backbone.history.start()

To have Backbone begin managing url changes (i.e. call default route, listen for routes, manage browser history, and deal with back button) you will need to call the Backbone.history.start(). The details for this method can be found in the section that talks about Backbone.History. For now, just realize that Backbone has to manually tell the browser to start the routing process. In the code below I setup a 'help' route but a change to the url is not monitored by Backbone until Backbone.history.start() is invoked.

7.3 - Backbone Initial Route

The initial route is the route that runs when Backbone.history.start() is invoked. The initial route is defined by sending the router an empty string route name signifying that this route should run by default (i.e. the root) when routing begins. In the code example below I define an initial route that is called after Backbone.history.start() is invoked.

Notes:

7.4 - Defining Routes

Up to this point the routing code examples have been defining routes using the route() method after the instantiation of a Backbone.Router() instance. To be comprensive, defining routes can occur not just by way of the route() method but in the following four ways (including the route() method):

1 - When Extending Backbone.Router Using routes Option:

2 - When Instantiating A Router Using routes Option:

3 - After An Instance Is Created Using The route() Method:

4 - During Initialization Of An Instance Using The route() Method:

7.5 - Running Routes With Wildcard Pathnames (aka splats or *)

A wildcard pathname can be used to indicate that any pathname will cause the route to run. For example, the route '*anyURL' will run anytime a url is changed.

And the route '/help/*' will run anytime the url changes to 'help/ANYTHING/ANYTHING'.

Notes:

7.6 - Passing Parameters From The URL To The Routes Callback Function

To identify url parts that should be considered parameters simply place a : in front of the name of the parameter contained in the route string. For example in the code below I pass the search route callback function, 4 parameters.

Notes:

7.7 - Using A Regular Expression With route() to Match A URL

The route() method optionally accepts as its first parameter a regular expression value instead of a string value. This can be pretty handy when you'd like to finely tune the matching of urls. In the code example below I demonstrate the matching of a route for a specific range of words.

7.8 - Routing Broadcasts A Built In 'route:CALLBACK-NAME' Event

When a route is match and the callback function is called a built in 'route' event, with a namespace referring to the name of the callback invoked is dispatched/broadcasted by Backbone. The on() or listenTo() event methods can be used to tap into the 'route:NAMEOFCALLBACKFUNCTION' event.

In the code below I am using on() or listenTo() to monitor and invoke additional callback functions when the search callback is invoked.

Notice that both on() or listenTo() pass along any url parameters.

Note:

7.9 - Manually Navigating Routes using navigate()

Up to this point I've mostly been invoking routes in the code example by changing the window.location.hash value in the browser, which Backbone is monitoring. By using the navigate() method we can accomplish the same thing. In the code below instead of update the hash value in the browser I am going to invoke the navigate() method passing it the route string to navigate too.

Notes:

8 - Backbone.History

The Backbone.Router is responsible for managing routes. The Backbone.history is part of routing and responsible for listening and responding to URL changes, including updates to the browsers history. The Backbone.History() constructor is instantiated by the Backbone library itself and an instance of History() is referenced from Backbone.history. The object that is placed at Backbone.history has one method named start(). Calling this method tells Backbone to begin listening for routes and to start managing the browser history. This method takes an options object with the following options and values.

Backbone.history.start({

/*Boolean, default is false, true means use pushState if available and fallback to hash paths if needed.*/
pushState: true,

/*Boolean, default is true, a value of false if pushState is true means that browsers that don't support pushState will do a traditional browser reload based on URL pathname.  If pushState is false and hashChange is false then url changes will do a traditional browser reload based on URL pathname. */
hashChange: false,

/*String, default is '' and backbone assumes you are serving if from the root url (i.e. /).  A value other than '' is a string path to the directory your application is being server from and the directory Backbone should consider the root directory. */ 
root:'',

/* Boolean, default is false, a true value means don't run the default route (i.e.''). */
silent: true

});

You'll likely never create a History() instance in your code, but keep in mind that you'll almost always need to invoke Backbone.history.start() in order to begin monitoring hashchange events and invoke route callbacks.

Notes:

9 - Backbone Utilities

9.1 - Restore and Create A Unique (i.e. custom namespace) Reference to Backbone with Backbone.noConflict()

When Backbone is being parsed by the browser, one of the first things Backbone will do is store a reference to the value contained at Backbone in the Global scope (i.e. typically at window.Backbone). This is because Backbone will overwrite or take over this namespace and wants to give the developer the option to restore the previous value used before the Backbone library is parsed. This is where the Backbone.noConflict() function comes into play. Calling it will return the Backbone namespace to the previous owner and then return a reference to the last parsed version of Backbone.js so you can create a new namespace for it. In the code below I create a fictitious Backbone version before Backbone.js is loaded

<script>

    this.Backbone = {
    "I'm the previous owner of Backbone in the global scope": true
    };

</script>

and use Backbone.noConflict() to make sure this version remains reference from Backbone while giving the last parsed version of Backbone a new namespace.

Notes:

9.2 - Assign the DOM & XHR library used by Backbone with Backbone.$

Backbone will internally and automatically use any version of jQuery, Zepto, or ender (in that order) found in the global scope for DOM manipulation and XHR (aka AJAX) by storing a reference to one of these libraries at Backbone.$.

If none of these libraries are found in the global scope, Backbone will attempt to use whatever value is found in the global scope for $, which could be undefined if you don't include one of the previously mentioned libraries or define your own library which uses $ as a namespace or alias.

One can also manually provide a value for Backbone.$ after Backbone has been parsed. In the code example below I demonstrate how Backbone.$ can be set by either method.

<!DOCTYPE html>
<html lang="en">
<body>
<!-- functional programming utility library -->
<script src="http://cdnjs.cloudflare.com/ajax/libs/lodash.js/1.3.1/lodash.underscore.js"></script>
<!-- needed for older browsers without window.JSON -->
<script src="http://cdnjs.cloudflare.com/ajax/libs/json2/20121008/json2.js"></script>
<!-- DOM & XHR utility -->
<script>
//Backbone will use this if jQuery, Zepto, and ender don't claim $
var $ = {version:'0.0.1',name:'MyReCreateTheWheelDOM&AJAXLib'}
</script>
<!-- Backbone.js -->
<script src="http://cdnjs.cloudflare.com/ajax/libs/backbone.js/1.0.0/backbone-min.js"></script>
<!-- DOM & XHR utility -->
<script>
//Or just Manually set it after Backbone is loaded
Backbone.$ = {version:'0.0.1',name:'MyReCreateTheWheelDOM&AJAXLib'}
</script>
</body>
</html>

10 - Conclusion

I've read and watched a great deal of tutorials on Backbone.js and none of them provided a comprehensive foundation (all the parts around the meat) for grokking the function of a Backbone.Model,Backbone.Collection, and Backbone.View. It was my intention to alleviate this gap of information with the contents of this article.

With a thorough Backbone backdrop set into place by this article we will move on to the main stage actors, Backbone.Model,Backbone.Collection, Backbone.View and Backbone.sync in the second part of this article.

Now Available: Part 2: Backbone.js Deconstructed