The SALT interpreter is similar to a graphical browser, except that it recognizes only voice commands (rather than mouse clicks or text input). Like graphical browsers, it completes forms on the page and submits information to the Web server. In speech-enabled Web applications, users speak commands to complete the form controls.
Note The SALT interpreter interprets only a subset of HTML, since much of HTML refers to GUI or visual elements. However, the presence of GUI-related tags does not introduce any errors.
Unlike Internet Explorer, the SALT interpreter does not allow the structure of the Document Object Model (DOM) to be altered, once it has been created for a document.
When Telephony Application Services (TAS) starts, it dynamically creates the required number of SALT interpreter instances. At least one SALT interpreter instance is required per caller.
When an instance of the SALT interpreter is initialized, it loads a Start page from a Web server. The Start page can be compared to a home page in other browsers. When the caller's session terminates, the SALT interpreter resets and reloads the Start page.
The following diagram illustrates how TAS and a SALT interpreter process a request for speech recognition.