Ajax and “Atlas” the new web way?

Web architecture has long depended on the stateless and fairly unintelligent serving of pages via HTTP requests. The user does something and the client sends an HTTP request for a new page. While the page is being delivered the user has to wait, usually looking at a partly completed page. A full HTTP request is needed, even if the requirement is only to validate a single item of data. One solution is to use client-side scripts, but for various reasons these have fallen out of favour. AJAX – Asynchronous JavaScript and XML – is the new model that overcomes these problems. Instead of loading a web page, the browser loads an Ajax engine, which is simply a client-side server written in JavaScript. The engine renders the interface and communicates with the remote server. Actions that would have resulted in an HTTP request call the Ajax engine instead. If the engine needs to talk to the server, then it can do it in a fast and more efficient way using XMLHttpRequest, utilising XML and XSLT for data exchange and manipulation. Although Ajax is relatively new, it is already used in some high profile web sites, including Google Groups, Suggest, and Amazon’s A9 search engine.

Ajax tools and utilities are currently a big growth area, and even Microsoft has an Ajax implementation, code-named “Atlas”. There are, however, concerns over the impact of a new client architecture in blurring the distinction between the systems layers – presentation, logic and business layers. There is also no consensus as to the effect on maintenance of introducing so much JavaScript. Indeed, there is the view that Ajax is just an attempt to get the mess that is JavaScript under control.