VoiceXML is an XML-based language suitable for use by any knowledgeable XML programmer. With the release of the VoiceXML 2.0 Specification, VoiceXML will increasingly be employed by organizations to voice-enable XML content. This session describes best practices for VoiceXML and includes a development scenario for rapid VoiceXML application development using current XML tools and nascent VoiceXML IDEs.

Attendees will receive an overview of the major components of a VoiceXML application. Join us for a discussion of the steps XML programmers should take to create VoiceXML, validate applications, and deploy a VoiceXML interface to existing systems within an organization. This session will also include a brief introduction to the necessary components of a voice development environment along with an overview of existing tools.

This presentation will guide the audience through a voice development environment and describe straightforward processes for adding VoiceXML output to an existing XML content workflow, including the possibility of transforming existing XML content directly into VoiceXML via XSLT. Changes to these types of generated voice interfaces will be covered -- one example is the substitution of VoiceXML session variables for session cookies.

The presenter will also touch on the different choices of output available to the end user: TTS has advanced considerably since the days of the "drunk Bavarian" -- and concatenated speech has nearly erased the difference between a completely pre-recorded prompt and a synthesized one.

In addition, this session will explore a number of valuable topics including current ASR engine technology and the differences between engine offerings from vendors including Nuance and SpeechWorks. Finally, the presenter will outline the implications of processing choices. While a VoiceXML platform itself can often do the majority of the processing, processing done on the server can reduce platform incompatibilities and debugging problems.