Welcome to 10 Projects

This is the companion site for the VoiceXML: 10 Projects to Voice Enable Your Website book by Mark Miller. Here you will find the most current scripts for each of the projects in the book, enhancements that have been suggested by people using the book, information on how to access the demos, and links to various resources.

You don't have to own the book in order to use the scripts; they're ready to plug and play. However, if you want to know how the scripts work, the book will be your best resource.

To get started, choose one of the projects on the menu at the right. If you are new to VoiceXML, proceed sequentially through the first five projects. You'll then have the basics for completing the rest of the projects. Experienced developers can choose from the later projects, but Project 4 on User Authentication might be interesting.

Thanks for participating in the projects. I look forward to your feedback.

Mark Miller

Projects

Project 1: Building Basic Pages

Beginning to use VoiceXML to build voice applications as an extension to your Web site is as simple as learning how to build your first Web page with HTML. VoiceXML has a predefined set of markup elements that can be applied to your online content for output. This first project will be used as a template for subsequent projects. It contains the basic building blocks that can be used as an underlying structure for expansion to production-level applications.

THE PROBLEM
Most Web developers don't have experience with voice applications. They need a step-by-step solution to get an initial project up and running.

THE SOLUTION
Create a basic template that can be used for developing larger applications. Building the template includes creation, debugging, and deployment of the template.

Resources
for this Project

Project files
Errata
Newsgroup
Enhancements
Demo information
About the book

Project 5: Templates and Multiple Documents

The previous projects each developed specific parts of a Web page and were added to a single document. When building a large Web site, however, it is much more efficient to break up the application into separate components and access them from a single base document. This project will take forms built in the first four projects and store them in separate documents, making it possible for various members of your development team to work on a single component. We will examine techniques for building generic subdialogs that can be used in multiple applications.

THE PROBLEM
As single-document applications develop into larger applications, it is difficult to maintain them and there is a greater chance of code duplication.

THE SOLUTION
Develop a set of common components for generic forms and functions common to most VoiceXML applications.

Resources
for this Project

Project files
Errata
Newsgroup
Enhancements
Demo information
About the book

Project 6: Telephone Dialer

This project builds a simple vocal telephone dialer that stores names and phone numbers and allows you to access them through one phone number. You access a phone number in your list by speaking the name of the person you want to call. The application will retrieve the phone number and dial it for you. All you have to do is remember the one phone number that accesses your contact list.

THE PROBLEM
Whenever you need to retrieve a phone number, whether at home, at work, or on the road, you have to find your address book, look up the phone number, and dial it. It is almost impossible to keep track of all telephone resources through one central source.

THE SOLUTION
Create a comprehensive telephone directory that can be accessed by voice through one toll-free telephone number.

Resources
for this Project

Project files
Errata
Newsgroup
Enhancements
Demo information
About the book

Project 2: Creating Links

One of the reasons the World Wide Web continues to grow is because it gives programmers the ability to create links to a wide array of resources. A "mom and pop" online bookstore can sell any book in print by linking to Amazon, Barnes and Noble, or Books-A-Million. VoiceXML has several ways to create links. In this project, we'll investigate the tricks and traps associated with vocal linking by creating an information resource for the Mars 7 Colony Visitor's Bureau.

THE PROBLEM
A VoiceXML application that has any kind of complexity will use multiple data sources and content resources for its output. How can the application access the data sources through one central document?

THE SOLUTION
Create complex application flow using the various linking mechanisms available in VoiceXML to access a variety of data sources, internally and externally.

Resources
for this Project

Project files
Errata
Newsgroup
Enhancements
Demo information
About the book

Project 3: Contact Us

Most Web sites have a Contact Us section, which displays various ways a user can contact the company or Web staff. I've always been intrigued with the possibility of using a robot like Hal from the movie 2001 for handling this type of mundane chore. Not only would it be able to give you basic responses to your contact requests, but it could also help do simple tasks like taking messages or transferring calls to a person. This project can be used to lay the groundwork for developing a semi-intelligent answering system that handles the time-consuming chores of your front office staff.

THE PROBLEM
Using a live receptionist to answer basic questions about your company or to take phone messages is a waste of personnel time and company resources.

THE SOLUTION
Access to basic contact information about your company through verbal output of an address can be handled with an automated system. The user has the choice of requesting company information, leaving voicemail, or speaking with a live receptionist.

Resources
for this Project

Project files
Errata
Newsgroup
Enhancements
Demo information
About the book

Project 4: User Authentication

In the early days of the World Wide Web, the network resembled a close-knit neighborhood. Everyone knew everyone else, free information was available, and anyone could participate. With the advent of e-commerce and online business, the game has changed. Private networks are using the resources of the Internet, allowing access to privileged information such as customer records, proprietary company data, and personalized account information. This project looks at several solutions for validating users through voice authentication procedures.

THE PROBLEM
Web sites need ways to block users from gaining access to private information and to log in users to refine the user's profile.

THE SOLUTION
Create a user authentication system, using multiple ways to confirm the user's identity.

Resources
for this Project

Project files
Errata
Newsgroup
Enhancements
Demo information
About the book

Project 7: Report Recorder

I lived in Hawaii for three years. Waves of nostalgia break over me every once in a while and I just have to know what the break is like at North Shore. www.surfline.com has Web cams and reporters set up at beaches around the world. The reporters log in and type a surf report each day for the Web site, but that doesn't do much good when all you have is a telephone. That's where the idea came from for this project. A special thanks is offered to Tim Chandler and his crew at Surfline.com for making the surf data available for this project.

THE PROBLEM
Multiple sources of information need to be recorded and stored on a daily basis. The people making the reports should be able to call in their reports over any telephone.

THE SOLUTION
A voice recorder that stores reports in .wav file format will allow the remote reporters to log in, make their reports, and have the reports accessible to users through a telephone call or a link on the Web site.

Resources
for this Project

Project files
Errata
Newsgroup
Enhancements
Demo information
About the book

Project 8: Quiz Generator

Telephone game quizzes and adventure hints are natural extensions to online games. The Quiz Generator project came about when I discovered a site on the Internet that contains over 18,000 sound files from the Star Trek series. I thought it would be fun to insert some of the character voices into a game format and have people guess which character was speaking. A special thank you goes to Dave Wythe at Star Trek in Sound and Vision for providing the sound files for this project.

THE PROBLEM
Most companies have multiple sets of data, stored in structured format, that need to be output in the same format. The data is hand-coded into a preset template, causing input errors and wasting valuable time. Once the data has been input into the template, it is no longer available for other applications to use.

THE SOLUTION
Isolate the data from the formatting. Store the data in a nonproprietary format, such as XML. Use XSLT as a generic formatting mechanism, outputting VoiceXML on the fly, while leaving the data source untouched.

Resources
for this Project

Project files
Errata
Newsgroup
Enhancements
Demo information
About the book

Project 9: San Francisco Arts Calendar

There are thousands of events, give or take a couple hundred, at any one time in the San Francisco Bay Area. The San Francisco Arts Organization, www.SFArts.org, has a site that coordinates listings of all of the events into disciplines, such as dance, music, and theater.

This project will use the database from the SFArts project and create a voice interface for accessing the calendar of events. A special thanks is offered to Larry Larson, site manager, for providing access to the San Francisco Arts Calendar data and for creating and debugging the Cold Fusion template.

THE PROBLEM
The general public wants access to a calendar of events for all types of performances in the San Francisco Bay Area.

THE SOLUTION
Create a voice interface to the database of existing events that is stored on the San Francisco Arts Organization site.

Resources
for this Project

Project files
Errata
Newsgroup
Enhancements
Demo information
About the book

Project 10: FAQ Engine

I met Kevin Dowd in 2001 at a Speech-Tek convention in New York. The display for his product, Brainhat, said that he had a VoiceXML-enabled speech engine that could carry on a conversation based upon context, not word recognition. After testing out the demo, I became a believer. The FAQ Engine project was created by me, Kevin, and his team at Brainhat. It demonstrates how to access a natural language processing engine through a VoiceXML interface.

THE PROBLEM
Most FAQ engines have preset definitions for a set of 10 most frequently asked questions. There is no dialog between the user and the information, just a question and a hard-coded answer.

THE SOLUTION
Build a voice-enabled FAQ engine that can infer relationships through context and handle basic natural language processing.

Resources
for this Project

Project files
Errata
Newsgroup
Enhancements
Demo information
About the book