VoiceXML in the Large

By Rob Marchand

Welcome to First Words, VoiceXML Review's column that teaches you about VoiceXML and how you can use it. We hope you enjoy the lesson.

Last month we had a quick look at how a VoiceXML document is structured. Interesting applications can be built with only a single static page, or a combination of a few static pages. But what if you're building a large application, with many possible user interactions, and access to dynamic data?

VoiceXML provides capabilities for developing large application. In addition, the web-centric model on which VoiceXML is based, allows the use of conventional web server technologies for generating dynamic content. Typically these involve templates that are processed by a web application server to insert dynamic content . For example, Java Server Pages containing VoiceXML templates, generate dynamic pages by using JavaBeans to generate the dynamic bits.

This month, then, we're going to introduce the elements of VoiceXML that support the building of large applications in these environments.

Application Structure

VoiceXML applications are comprised of a collection of documents. From our previous column, you will recall that a document is the equivalent of an HTML page, and encapsulates one or more dialogs. Execution of a dialog typically involves presentation of information to the caller, along with collection of input from that caller. Transition from one dialog to another is controlled by the currently executing dialog. Transition occurs via the <goto> or <submit> tags.

The simplest VoiceXML application (see Figure 1) is simply going to be a set of VoiceXML pages, returned by a web server to the VoiceXML platform. These pages are interpreted to control the interaction with the user.

Figure 1: Static VoiceXML Application

The application is entirely analogous to a set of HTML pages delivered by a web server. Transitions are controlled by the <goto> tag (which is executed in VoiceXML, as compared with being 'clicked' by the user in a visual web browser).

Dynamic applications function in much the same way, with the exception that some or all of the pages are generated dynamically, by a server-based technology such as JSP/JavaBeans, ASP/ActiveX, ColdFusion, or scripting languages such as perl. An overview of this model is shown below.

Figure 2: Dynamic VoiceXML Application

VoiceXML defines an application as a collection of documents, possibly tied together with an application root document. The application root document provides a mechanism for retaining state among a related set of VoiceXML documents. A VoiceXML refers to its application root document using an attribute in the <vxml> tag. This means that the event handlers, variables, and global grammars defined in the root document are available to the subdocument for use. This structure is shown below:

Figure 3: Application Level Information

So while VoiceXML pages do not retain state across <goto> and <submit>, the application root document allows the definition and management of data and behavior that is persistent across VoiceXML pages. This allows both static and dynamically generated pages to manage global data, and to provide consistent handling of events, and access to global (and consistent) behavior across an entire application.

back to the top

Copyright © 2001 VoiceXML Forum. All rights reserved.
The VoiceXML Forum is a program of the
IEEE Industry Standards and Technology Organization (IEEE-ISTO).