Skip to content
HTML/CSS/JS dependency and build task free boilerplate
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.



The boilerplate is a sample project to build MVC javascript websites, either in single or multipage formats with the least overhead possible. By the use of extremely small micro-libraries, basic javascript functions and structures can remain the style and structure of the website.

Table of Contents


Instead of relying on each member of your team to add and update their .git/hooks/ files individually, the hooks are instead kept in the repo itself.

This means that you can add as many hooks directly into your repo as you like, and you can also add custom local hooks that can be individually ignored as well.

To set up the hooks in your local repository, run the following executable:

$ hooks/hooks-installer

Any hook defined in the hooks/ folder will then act as if it were in .git/hooks/. A pre-commit hook has been included by default. Upon any commit, the pre-commit hook will auto generate a file and append its contents at the end of this file. The contents of the list is generated by searching all javascript code for instances of //TODO:.


The overall objective of the javascript used in this boilerplate is to do away with the need for 3rd party libraries like jQuery. Instead, small libraries are created with as little dependency as possible. These can be chained together and use one another by way of the NS (NameSpace) library. By eliminating jQuery and other massive libraries, the focus can be back on writing pure, vanilla javascript code. There is no other framework to learn or adapt to. In fact, the MVC pattern used below in the App section can be completely removed and replaced by a single page file if that better suits your needs. Everything here is designed to be as modular as it can with the exception of NS.js, which serves as the dependency loader.



NS.js (Namespace) manages the loading and access of dependencies across the application.


To define a class, library, or module, use the NS function. NS will make the necessary ajax request to load the any dependencies. When complete, it will fire the callback function. If the callback function returns any value, it will be stored in a global object matching the id string.

// NS (id, libs, callback, scope);
NS ( '', ['', 'another.lib.Loading'], callback, scope);
// creates a new object at

There are three required parameters in the NS method:

  • id : If you are defining a class, this is where you name it. If your callback method has a return value it will be stored in an object at this path/name. Dot syntax works here (e.g., ''). If you are not creating a class, you should still name the closure with an id as a means of debugging.
  • libs : An array of other namespaced objects to load as dependencies. An empty array is required if there are no dependencies.
  • callback : A function to call when all dependencies have been loaded. If this returns anything other than null, it will be stored in an object matching the id string.
  • scope (optional) : If you want your callback function to operate in a scope other than window, define it here.

The NS function will attempt to identify infinite recursion in dependencies. If this is encountered an error will be thrown and execution halted.

This method was added as a result of xHR synchronous loading being deprecated in the latest JS spec.


To use another class or library into the current context, use the NS.use method. You can directly reference the lib by its full dot-path at any time. This method is provided as a convenience.

var LocalCopyOfLib = NS.use ("");

Setting the NS.debug property to true will enable console logs to help debug dependency or load issues you might encounter.

NS.debug = true;

Setting the NS.baseURL property will allow you to change the root path of your external javascript libraries. By default the path is /. Be sure to include a trailing slash.

NS.baseURL = 'js/';

As the NS class is responsible for loading external resources, it is also capable of creating a browser-appropriate XMLHttpRequest or ActiveX object. This method calls a factory pattern returning the appropriate object and has been made externally available as a convenience method.

var httprequest = NS.createXMLHTTPObject();


Ajax.js is a convenience wrapper for the loading of all ajax requests. It contains only a single constructor method.

new Ajax()

The Ajax class takes 3 or 4 parameters: url, callback, error, [post data].

var getExample = new Ajax ( url, callbackFunction, errorFunction );
var postExample = new Ajax ( url, callbackFunction, errorFunction, postData );

If a readyState of 4 and request status of 200 are received, your callback function will be executed. Otherwise the error function will be called.

Note: calling Ajax without the new keyword will throw errors.


Analytics.js is a skeletal structure designed to wrap and centralize your website analytics calls. By default it is configured to use Google Analytics. In its present form it does little but wrap that functionality into a class structure.

This class shines best when you have other types of metrics being collected. Combining these calls into a single class saves time and makes the code more readable.

The Analytics.js class automatically hijacks exit links and adds event tracking.

new Analytics()

Initialize your analytics class by passing it your Google Analytics ID.

var analytics = new Analytics ( 'UA-971204-A' );

The Analytics class automatically interupts all links that exit your domain to immediately track an exit event before navigation.


Track events in a convenient way that doesn't interfere with bounce rates or timing.

analytics.trackEvent( category, action, label, value );

Both the label and value parameters are optional.


Timer events are easily fired off with this helper method.

analytics.trackTime( category, variable, value, label );

Value and label are optional parameters.


Bidirectional data binding is made available through the Bind class. Instantiating an instance of this class requires a single id parameter. This ID will connect to one or more data-bind-* properties in the DOM.

var bindsample = new Bind('bindsample');
// Relates to elements with the attribute "data-bind-bindsample"

Any change events fired by that element will automatically update the data in the Bind instance's attributes list. Likewise, any change to the Bind instance's attributes will propagate to the DOM.

Setting or getting an instance's attributes is done via the set and get methods.

bindsample.set('propname', 'samplevalue');

The above example will update the value or innerHTML of any elements with the appropriate attribute, like the following example:

	<div data-bind-bindsample="propname"></div>

Properties values are updated or inserted into these DOM elements via innerHTML, or value if they are input, textarea, or select types.


Some events fire far too often to be properly usable. The scroll event is notorious for this. Debounce is a limiter for events. You can indicate how often any given event should be allowed to fire and the class will filter the rest.

// Debounce ( callback, fireRate, fire_at_start, fire_at_end )
var debounceScroll = Debounce(hasScrolled, 66, true, true);
window.addEventListener("scroll", debounceScroll, false);

In the above example, the scroll event will be limited to firing once every 66ms. An event will be triggered immediately, and an extra after the very last polling.


DOM.js provides some basic cross-browser methods for retrieving and creating DOM elements.


Pass a string of HTML to this method to create the elements quickly by way of createDocumentFragment. Don't forget to attach the result where you want it, though.

var el = DOM.create('<p>hello world!</p>');

One of the most obnoxious limitations of working without jQuery is the inability to quickly find elements in the DOM. This method helps sort that out.

var elementByID = DOM.find('#someid');
var elementsByClass = DOM.find('.someclass');
var elementsByNode = DOM.find('li' );
var elementsByContext DOM.find('.someclass', elementForContext);

To best support older browsers, this method does not try to handle child selectors or anything too fancy. It's faster just to make separate queries and use the result as the context of the next.

Also, IE8 does not support querySelectorAll. Certain queries may generate a warning about this fact in the console.


Instead of finding and returning a DOM element, you can remove it with this method. You can use the remove method with the same limitations and examples as the find method, or you can remove an element directly by passing the DOM reference.

// Remove a direct reference to a DOM element
var elementByID = DOM.find('#someid');

// Remove combined with the built-in find

Removes a class from a DOM element.

// Remove a class by direct reference to a DOM element
var elementByID = DOM.find('#someid');
DOM.removeClass(elementByID, 'someclass');

// Remove a class with the built-in find
DOM.removeClass('#someid', 'someclass');

Adds a class to a DOM element.

// Add a class by direct reference to a DOM element
var elementByID = DOM.find('#someid');
DOM.addClass(elementByID, 'someclass');

// Add a class with built-in find
DOM.addClass('#someid', 'someclass');

Returns a bool indicating whether the DOM object contains a certain class.

var hasClass = DOM.hasClass(element_name_id_or_reference, 'someclass');

Adds or removes a class on a DOM object.

DOM.toggleClass(element_name_id_or_reference, 'someclass');


Delegate.js is a very simple wrapper that enables a global function called Delegate, which proxies the context of this in a function. When using class-based structures, it's used to maintain class scope in event listener callbacks, especially on the DOM.

// create a delegate function maintaining scope
var c = Delegate(callbackFunc, this);

// examples of using that function
Event.subscribe ('SOME_EVENT', c);
Event.unsubscribe ('SOME_EVENT', c);


Draw.js wraps basic drawing helpers.

Draw.line ()

The line method allows for the drawing of very simple lines by creating DOM objects with the appropriate CSS transforms.

// Draw.line (x1, y1, x2, y2)
var line1 = Draw.line (100,200,100,400);

Simply style the line and you're good to go. Here's a sample styling:

position: absolute;
height: 0px;
border-width: 1px 0px 0px 0px;
border-style: solid;
border-color: #666;
transform: translateZ(1px);
-webkit-backface-visibility: hidden;
-moz-backface-visibility: hidden;
-ms-backface-visibility: hidden;
-o-backface-visibility: hidden;
backface-visibility: hidden;


Easing.js adds a few easing methods onto the Math object.


Events.js is a simple subscriber that allows for a basic global messaging system. The main methods are subscribe, unsubscribe, and trigger. There are also various logging functions for debugging purposes. This Events class allows for the observer pattern which in turn supports the MVC structure of the boilerplate.


The subscribe method supports 3 parameters: eventName, callback, [priority].

Events.subscribe('EVENT_NAME', callbackFunc, 10);

By default, events are not proxied to preserve scope. You will likely want to use Delegate when using this within a class.


The opposite of the subscribe method, this will unbind your event listener. Only the event name and callback are required. When unsubscribing an event with a callback that was created using a Delegate, be sure to save the reference to the proxied callback for unbinding.

Events.unsubscribe('EVENT_NAME', callbackFunc);

Triggering an event is simple with the trigger method. Three arguments are supported: eventName, [data, context].

Events.trigger('EVENT_NAME', [arr, of, data]);
Events.trigger('EVENT_NAME', [arr, of, data], this);

The data property must be an array. If it isn't, it will be converted to an array.

The context value will set the scope of this in the callback function. Please note this could have strange consequences if you've proxied the callback function.


JSONP.js is a microlibrary to support basic JSONP requests. Ajax can't be used to handle JSON requests across domains, but JSONP can get the job done. The implementation is quite different, and this lib sorts it all out in a cross-browser way. The usage below is self explanatory.

jsonp('', callback, error);


It can be a pain to detect which prefix to use in CSS when developing without build tools.

Prefix.hasStyle ()

Determines if the browser supports one of the css selectors in the list. Returns bool.

var hasTransition = Prefix.hasStyle('transition WebkitTransition MozTransition MsTransition OTransition');
Prefix.getStyle ()

Returns the browser supported css selector in the list provided. Returns matching string or null.

var transition = Prefix.getStyle('transition WebkitTransition MozTransition MsTransition OTransition');


ScrollTo.js provides a way to smoothly animate the page scroll to a new position with custom easing and callback methods.

//ScrollTo(to, callback, duration, easing);
ScrollTo(0); // scrolls to top of page using default easing
ScrollTo(, doneAnimating); // Scrolls to bounding rect of object then calls callback
ScrollTo(0, null, customEasingMethod); // define your own easing method


Storage.js is a localstorage wrapper. It supports JSON processing of objects being stored and retrieved. This static class has a getter and setter property.

var value = Storage.get('someid');
Storage.set('someid', someValue);


Storage.js is a basic templating engine in the style of Handlebars. The syntax is a bit more strict, requiring a matching close tag with id. There is no control logic, but looping is supported.

var Template            = NS.use('lib.Template');
var templateExample     = '<p id="{{ id }}">{{ content.text }}</p> <ul> {{#each item}} <li>{{ label }}</li> {{/each item}} </ul>';
var templateData        = { "id": "templateTest",
							"content" : {
							  "text" : "template text example"
							"item" : [
							  { "label" : "item 1" },
							  { "label" : "item 2" },
							  { "label" : "item 3" },
							  { "label" : "item 4" },
							  { "label" : "item 5" },
							  { "label" : "item 6" }
var renderedTemplate    = Template(templateExample, templateData);
var renderedHTML        = DOM.create(renderedTemplate);


The design of the application logic follows the MVC pattern by way of Observer patterns. A subscriber system allows for global binding and triggering of events to control application flow. In a multi-page application, the page source (js/page1.js for example) will load the appropriate Models, Views, and Controllers for the given page's content. Components can be reused or shared between pages to minimize asset load at runtime and facilitate good OOP principals.

If you prefer to use a different presentation pattern than MVC, the libraries available in this boilerplate should allow you to easily restructure the app to your needs. I recommend reviewing the excellent blog post by Manoj Jaggavarapu on the subject.

Like all library dependencies, each application component makes use of NS.js to create a namespace for the class and to handle dependency loading.


The model controls the business logic of you application. Business logic can be a vague term when it come to apps that are primarily front-end. For years I was tripped up by this and ended up with massive Views as a result. Take heed from my mistake. If your app doesn't have business logic to speak of, consider this a ViewModel. It's flexible and can do whatever you'd like.

In the SampleModel.js file, an example of a constructor, public method, getter/setter, and internal methods are available. The class demonstrates a very basic example of a state machine, via ApplicationMode, which may be useful. Methods are commented with docstring-esque syntax, which may help support some code editors with better code inspection support.

The sample model also gives an example of external data loading via an example JSON file loaded via ajax.


The Views of your application are in charge of the DOM. They're created by passing a reference to the part of the DOM they will control. They can then create, replace, update, and delete the visible interface. They also handle any bindable interactions (e.g., buttons, textfields, swipes).

When an interaction takes place in the View, it dispatches Events to inform the rest of the application. The View affects no change in the application on its own.

The View will listen to Global data events relevant to any data it is displaying in order to update itself. Application state changes, list updating, pagination, are all examples of Events that might be subscribed in the view.

You may also make use of the Bind class for bidirectional data binding. See the section above for details.

By default there are no dependency libraries in this boilerplate for accessing or manipulating the DOM. JQuery, in particular, is overly large and counter to the philosophy of this tool.


Controllers listen for events from the views and control specific actions by manipulating the Model. Your controllers will have a reference to one or more Models in your application and will call public methods on those Models to make changes.

In the SampleController.js file, the example is subscribed to a UI Event. When that event fires, the controller calls a sampleMethod on the Model. In rare cases, Controllers may also fire events of their own, listened to by other controllers. Generally speaking your controller will listen for Events from the Views and call direct methods on Models.


This boilerplate is designed to use open source technologies, including PHP for backend control. There is very little standard backend as part of the core boilerplate, but there are a few nice utilities.

Adaptive Images

The Adaptive Images utility automatically sizes your imagery to minimize bandwidth on smaller devices. By measuring the width of the screen/client as the page loads, the backend script automatically resizes your image to the appropriate breakpoint. These resized images are saved in a cache folder to minimize processing load on the server, and are automatically served to the client by way of Apache MOD_REWRITE directives in the .htaccess file (see below). The GD lib is used (normally a default module installed with PHP) to support the resizing.

To set up adaptive images, all you'll need to do is tweak the $resolutions variable inside the adaptive-images.php file in the web root. Match these resolutions to your CSS breakpoints and everything should just work. If you encounter any issues, make sure that the images-cache folder (created automatically) has forgiving write permissions in order to create the resized images. Also, be sure that you are taking advantage of the included .htaccess file. It is hidden by default in many operating systems and may be easily overlooked by FTP clients.

Some small customizations have been made to the adaptive-images code in both PHP and javascript to improve performance.



The included .htaccess file performs a number of desirable tasks, including the following:

  • Rewrites all directory URLs with trailing /'s (for SEO)
  • Supports Adaptive Images (see above)
  • GZip compresses many common MIME-types
  • Sets Access-Control-Allow-Origin directives
  • Defines caching durations and access by type
  • Adds Cache-Control headers


The robots.txt file gives instruction to search engines crawling the site as to which content is appropriate to show in search results. By default a number of basic directories are excluded, most notably the images folder.

Finally a reference is made to the sitemap file, used most commonly as instructions to bots on not only what to spider, but how often to check back.


An example sitemap file is included here illustrating how to list the pages of your site, rank them via priority, and set their frequency of change. It is much easier to use a sitemap generator than update this by hand.


In the style of robots.txt, the humans.txt file provides information to... humans! There is very little to the standard formatting for this file, but it should be straight forward and easy to understand for any human that stumbles upon it. The format in this project losely follows that of HTML5Boilerplate.


A sample favicon.ico file has been included mainly so that it's not forgotten. Hopefully a big black box will serve as a nice reminder to put something meaningful together.

Anything below this line will be overwritten upon commit. See the section on Hooks, above, for more details.


You can’t perform that action at this time.