batman.js is a stateful MVC framework based on Rails conventions used for building client-side single-page applications. That’s a lot of buzzwords; let’s break it down.
1. Model-View-Controller Architecture
Model-View-Controller (MVC) is a common way of dividing the responsibilities of your application. There are three main parts:
- Model: Represents the data of the system, possibly coming from a database or web API.
- View: Interacts with the user by displaying the layout and data, and responding to user input.
- Controller: Handles communication within the app by requesting data from the model and rendering specific views with that data.
Different frameworks and systems and languages will all implement MVC in slightly different ways, but the overall concepts are similar. batman.js borrows its conceptual MVC implementation heavily from both Rails and Cocoa, which makes it simple to understand if you’re already familiar with one of them.
Here’s how the key components work in batman.js:
Models are still responsible for fetching and saving all your data, but since we’re a client side app, they’ll load it via AJAX from your server. It’s worth noting that batman.js is more or less backend-agnostic. It doesn’t really care what type of server you’re using, as long as it responds to common REST API’s with JSON (but you should still consider Rails for maximum easiness).
Like in Rails, a controller action gets called whenever a specific route gets triggered. For example, if a user accesses
new method of your
ProductsController will automatically be invoked. All you have to do is implement the correctly named method, load the data that the view will need, and then tell batman.js which HTML file to render. Most of this is handled implicitly for you, as long as you’re following the batman.js naming conventions.
Unlike in Rails, batman.js makes a distinction between HTML files and views. HTML files are the main way to get anything on screen. Developers and designers can both build vanilla HTML files and add extra
data-* attributes to tell batman.js to bind a particular node to a particular piece of data. There’s no extra templating language to learn, just a set of attributes.
Views in batman.js render a specific HTML file to the DOM, prepare presentation data for that HTML file’s bindings, and listen for user interaction with the parts they rendered. Most of the time, a view object will be automatically instantiated for you based on the controller action that is being rendered, but you can also create your own view classes to encapsulate pieces of view logic, e.g. a
MapView for displaying a map. The view subsystem of batman.js is actually very similar to that of Cocoa’s, so you get many of the same great API’s that are available to Cocoa developers.
2. Stateful Single-Page Application
One thing to keep in mind is that while a number of the API’s may look very similar to Rails, your batman.js application is operating in an inherently different environment. When Rails receives a request, it starts the application, spawns a controller, calls the action, hits the database, returns the HTML, and then shuts the application back down. We say Rails is stateless because the only state persisted through requests is in the database.
batman.js has no concept of requests. The app is loaded once when the page is loaded, and then starts listening to events such as the URL changing or the user clicking around the application. The framework maps a number of these events to their Rails equivalent (such as a URL change event causes a controller action to be dispatched), but you do need to keep in mind that you are operating in a perpetual and non-ephemeral enviroment. Every variable that you assign to your controller is going to stay on that controller until it gets reassigned. Similarly, every view is going to stick around until you remove it from the view hierarchy.
We say batman.js is stateful because the entire app is persisted in memory until the user navigates away from the page. Keep this in mind when building your client-side application!