| VoiceXML (VXML) is the W3C's standard XML | | | | Typically, HTTP is used as the transport protocol |
| format for specifying interactive voice dialogues | | | | for fetching VoiceXML pages. Some applications |
| between a human and a computer. It is fully | | | | may use static VoiceXML pages, while others rely |
| analogous to HTML, and brings the same | | | | on dynamic VoiceXML page generation using an |
| advantages of web application development and | | | | application server like Tomcat, Weblogic, IIS, or |
| deployment to voice applications that HTML brings | | | | WebSphere. In a well-architected web application, |
| to visual applications. Just as HTML documents are | | | | the voice interface and the visual interface share |
| interpreted by a visual web browser, VoiceXML | | | | the same back-end business logic. |
| documents are interpreted by a voice browser. A | | | | Historically, VoiceXML platform vendors have |
| common architecture is to deploy banks of voice | | | | implemented the standard in different ways, and |
| browsers attached to the public switched | | | | added proprietary features. But the VoiceXML 2.0 |
| telephone network (PSTN) so that users can | | | | standard, adopted as a W3C Recommendation 16 |
| simply pick up a phone to interact with voice | | | | March 2004, clarifies most areas of difference, |
| applications. | | | | and vendors are going through a rigorous |
| There are already thousands of commercial | | | | conformance testing process set up by the |
| VoiceXML applications deployed, processing many | | | | VoiceXML Forum, the industry group promoting |
| millions of calls per day. These applications | | | | the use of the standard. |
| perform a huge variety of services, including | | | | Two closely related W3C standards used with |
| order inquiry, package tracking, driving directions, | | | | VoiceXML are the Speech Synthesis Markup |
| emergency notification, wake-up, flight tracking, | | | | Language (SSML) and the Speech Recognition |
| voice access to email, customer relationship | | | | Grammar Specification (SRGS). SSML is used to |
| management, prescription refilling, audio | | | | decorate textual prompts with information on |
| newsmagazines, voice dialing, and real-estate | | | | how best to render them in synthetic speech, for |
| information. They serve all industries, and range in | | | | example which speech synthesizer voice to use, |
| size all the way up to massive national directory | | | | and when to speak louder. SRGS is used to tell |
| assistance applications. | | | | the speech recognizer what sentence patterns it |
| VoiceXML has tags that instruct the voice | | | | should expect to hear. |
| browser to provide speech synthesis, automatic | | | | The Call Control eXtensible Markup Language |
| speech recognition, dialog management, and | | | | (CCXML) is a complementary W3C standard. A |
| soundfile playback. The following is an example of | | | | CCXML interpreter is used on some VoiceXML |
| a VoiceXML document: | | | | platforms to handle the initial call setup between |
| <?xml version="1.0"?>> | | | | the caller and the voice browser, and to provide |
|
| | | | telephony services like call transfer and disconnect |
| When interpreted by a VoiceXML interpreter this | | | | for the voice browser. CCXML is also very useful |
| will output "Hello world" with synthesized speech. | | | | in non-VoiceXML contexts. |