So, why should one choose a client-side JavaScript MVC framework over other server-side frameworks, such as Rails or Asp.Net MVC? Typically, the controller methods of any server-side MVC framework return views (that is, a fully formed HTML). However, there are cases when a view needs some data through Ajax calls (for obvious reasons). In such cases, the question arises as to which controller (method) should send this data. This leads to complexities on two fronts, which are described here:
The view becomes complex: In this case, you have to understand not only the part of the view that is generated by the server, but also all the AJAX interactions happening on the view. Then, you have to decide where to include the JavaScript code related to the view—in the same view in script tags or in a separate JavaScript file. Often, a server-side MVC framework uses a different template engine to build the HTML. This problem can be mitigated using a template engine such as Handlebars (http://handlebarsjs.com/).
The controller becomes complex: Some of the controller methods return complete views, while others return data. If you are using a client-side MV* framework, then the server-side controllers are API controllers, which only send data to the client. It is up to the client to display the data in whichever way it pleases. This makes the controllers simpler.
Similarly, the client-side MV* framework itself gives you some well-defined mechanisms to organize your code as per the MVC paradigm (or whatever convention the framework wants you to follow). So, the view code also becomes simpler and organized. Moreover, every interaction of the view with the server happens through Ajax calls. This too brings uniformity of communication.