With all of the buzz of late around AJAX (an acronym first coined by Jesse James Garrett to stand for Asynchronous JavaScript and XML), it was only a matter of time before the discussion began about using this approach in VoiceXML applications. There is an interesting article in the new VoiceXML review covering this topic.
The approach described in this article makes use of the new element that is part of the VoiceXML 2.1 specification. I'm not clear on whether this approach is the same as AJAX, but it is indeed powerful.
AJAX applications rely on a JavaScript object called “XMLHttpRequest“ that allows a web interface to communicate with a backend server without transitioning to new page. More traditional web applications submit user input to, or request data from, a backend server using the HTTP GET and POST methods. The transmission of this data and the subsequent transition to a new page can sometimes be perceived as disjointed and inelegant by users. AJAX allows web interfaces to fetch or submit data without transitioning to a new page, making them feel smoother (I guess).
The VoiceXML 2.1 [ element](http://www.w3.org/TR/2005/CR-voicexml21-20050613/#sec-data) doesn't work exactly like this. It allows a developer to designate an XML document to be fetched by a VoiceXML interpreter as part of a dialog. It then exposes that document via the [JavaScript binding to the Document Object Model](http://www.w3.org/TR/DOM-Level-2-Core/ecma-script-binding.html). Its "AJAX-like" in the sense that it obviates the need for a page transition to render the data in the XML document, but there isn't any asynchronous communication with a backend server through an XMLHttpRequest object.
Given the recent hype, I can understand the desire to characterize the functionality provided by the new 2.1 spec as emulating AJAX. But I think it’s important to point out that callers to a VoiceXML application probably don’t perceive page transitions in the same way that users of a visual web application do. There are ways to manage user reactions to these transitions within VoiceXML and users need to be aware of how to use them effectively.