In February 1999, IBM alphaWorks Laboratory announced SpeechML -"a[n XML] markup language for building distributed network-based conversational applications. SpeechML is an XML application, meaning that it is defined as a set of XML tags. The primary elements are <page>, <body> <menu>, and <form>. Pages group other SpeechML elements, and serve as the top-level element for a SpeechML document. A body element may be plain text, or it may marked up to improve the text-to-speech synthesis using the Java Speech Markup Language (JSML)."
The new package from IBM "describes Speech Markup Language, or SpeechML. It also contains a conversational browser that implements SpeechML, and some demonstration SpeechML applications. Just as HTML can be used as a markup language for building network-based visual applications, SpeechML is a language for building network-based conversational applications. A conversational application is an application that interacts with the user through spoken input and output. A network-based application refers to one in which the elements of the conversation that define spoken output and input - SpeechML documents - may be obtained over the network. SpeechML could be used to enable conversational access from (for example) a car, a telephone, a PDA, or a desktop PC, to information sources and applications anywhere on the Internet. Potential applications include the same range of services and information now available visually through HTML. Use of a markup language to describe the spoken interaction with the user has a number of advantages: it makes possible network-based applications; it provides a seamless transition from one application to another; and it provides an easy-to-use tool to describe conversational dialogs." Members of the IBM development team include Bruce Lucas, Jennifer Lai, Jung-Mu Tang, Paul Chou, Paul Moskowitz, and Steven De Gennaro.
For a related XML technology, compare Sun Microsystems' "Java Speech Markup Language (JSML)." Differences are said to be that "1) JSML only supports spoken output, while SpeechML supports a conversation by describing both spoken output and spoken input; 2) the SpeechML design (including IBM's own browser included with the SpeechML package) makes use of Java Speech Markup Language (JSML) and Java Speech Grammar Format (JSGF) from the Java Speech API."
References:
[February 18, 1999] "IBM Offers Speech Extension to XML." By Rebecca Sykes. In InfoWorld (February 17, 1999). "IBM on Wednesday announced Speech Markup Language (SpeechML), which provides an open framework to add speech capabilities to Web-based applications. SpeechML is based on Extensible Markup Language, which is a specification for formatting data on Web pages. SpeechML will be considered by the World Wide Web Consortium as the standard for speech markup language, according to a statement from IBM. SpeechML is designed to let a Web site developer use 'tags' to add interactive speech capability to his or her Web site without being an expert in speech technology, the statement said. Tags are the way the developer marks certain Web site content as content that is to be spoken by an application, such as reading a menu to a Web site visitor, IBM said."
Related technology: Java Speech Markup Language (JSML)
Related technology: Motorola VoxML - "The VoxML markup language for voice applications allows developers to simply and easily add speech interfaces to their Web applications or content."
Compare: "Speech Synthesis Markup Language Specification for the Speech Interface Framework."
