November/December 2001

BY MARC ROBINS

Top of the news as we go to press is the announcement of a brand new consortium, the SALT Forum, which aims to extend existing Web markup languages, such as HTML, xHTML, and XML, to provide multimodal and telephony-enabled access to information, applications, and Web services from PCs, phones, tablet PCs, and wireless-enabled PDAs.

A diverse group of companies, including Cisco, Comverse, Intel, Microsoft, Philips Speech Processing, and Speechworks International, have joined forces to develop Speech Application Language Tags (SALT) into a new royalty-free, platform-independent standard designed to add speech recognition and synthesis and telephony capabilities to HTML/XHTML-based applications, and make multimodal and telephony-enabled applications and services faster and easier to create, deploy, and use. The multimodal access capabilities envisioned will enable users to interact with an application in a variety of ways: They would be able to input data using speech and/or a keyboard, keypad, mouse, or stylus, and produce data as synthesized speech, audio, plain text, motion video, and/or graphics.

THE RATIONALE
According to the SALT Forum, there are three major challenges that SALT will help address: 

1. Input on wireless devices. Wireless devices are becoming pervasive but lack of a natural input mechanism hinders adoption as well as application development on these devices. The SALT Forum believes speech is a compelling solution to the input problem, and SALT will standardize how speech input and output will work for Web applications on those devices. 

2. Speech-enabled application development. Speech-enabled application development is still a difficult task, and not yet the domain of most application developers. By starting with HTML/xHTMLs object and eventing methods and script, SALT will introduce speech to the Web developer and create a new class of tools to simplify authoring of some of these applications. 

3. Telephone users. There are 1.6 billion telephones in the world, but only a relatively small fraction of Web applications and services are reachable by phone. By enabling a tight integration between existing Web browser, server, and network infrastructure and speech technology, SALT will allow many more Web sites to be reachable through telephones. 

VALUE PROPOSITIONS
The SALT Forum, like most standards-bearing organizations, seeks to make the new language extensions the development solution of choice for developers intent on creating new Web-based communications applications. The value propositions for SALT go something like this: 

The Forum envisions that end users will be able to use SALT-based applications anytime, anywhere, and from any device using speech, text, or graphical interfaces independently or at the same time. They see developers seamlessly embedding speech enhancements into existing HTML, xHTML, and XML pages, using languages, technologies, and toolkits they have been familiar with for years. They promise that businesses will realize reductions in cost and complexity from the ability to offer common Web-based applications across multiple presentation media, and also save by leveraging their existing Web investments and expertise to eliminate the need to create discrete applications for each type of output. 

Finally, and most importantly for readers of this magazine, the Forum sees service providers deploying a broad range of applications that enable the widest range of services, offering new business opportunities and revenue streams that better serve both consumers and business customers. 

SALT is designed as a lightweight set of XML elements that enhance existing markup languages with a speech interface. SALT can be used equally effectively with all the flavors of HTML, or with any other SGML-derived markup. SALT does not define a new programming model; instead, it reuses the existing Web execution model so that the same application code can be shared across modalities. 

As a result, SALT can be used to write speech applications for a variety of scenarios. Multimodal applications can be authored by adding SALT to a visual HTML page, for rendering both on small devices such as PDAs and smart phones, and on richer clients such as the desktop. Using the SALT features for DTMF input and telephony call control, voice-only applications can be written for scenarios in which a visual display is unavailable, such as the telephone. This allows developers to reuse the code for processing user input. The application logic remains the same across scenarios: The underlying application does not need to know whether the information was obtained by speech or text input.

The main programming elements of SALT are: 

<prompt ...> for configuring the speech synthesizer and playing out prompts;

<reco ...> for configuring the speech recognizer, executing recognition, and handling recognition events;

<grammar ...> for specifying input grammar resources; and

<bind ...> for processing recognition results into the page.

Marc Robins is Vice President of Publications, Associate Group Publisher, and Group Editorial Director for Technology Marketing Corporation. His Change Agent column appears in each issue of Communications ASPTM magazine. Marc appreciates your feedback, and may be reached via e-mail at mrobins@tmcnet.com.

[ Return To The November/December 2001 Table Of Contents ]