Handling requests RESTfully?

Question

There is a push in modern scripting frameworks to use RESTful request routing instead of a one-size-fits-all URL path routing.

respond('GET',    '/posts',       callback);
respond('POST',   '/post/create', callback);
respond('PUT',    '/post/[i:id]', callback);
respond('DELETE', '/post/[i:id]', callback);

vs

respond('/posts',       callback);
respond('/post/create', callback);
respond('/post/[i:id]', callback);
respond('/post/[i:id]', callback);

I must confess that I am still using the old-and-busted method where the resource is accessible at the given URL regardless of the request type.

Is there anything I'm missing by using the old format?

Answer 1

In my view what you are missing on, is the inherent architecture of a RESTful API. From an implementation point of view, the separation of request, resource and callback function, that REST imposes, encourages you to use an architecture which, by design, is clear and understandable. Separation of concepts in distinct parts is very important, especially when dealing with security-sensitive applications like web APIs. It reduces the complexity of development, testing and maintentance, and thus makes them quite easier to handle.