- From: Dan Burnett <dburnett@voxeo.com>
- Date: Thu, 7 Apr 2011 08:11:07 -0400
- To: Cesar Castello Branco <Cesar@CastelloBranco.eng.br>
- Cc: public-xg-htmlspeech@w3.org
Speaking as one of the authors of VoiceXML 2.0/2.1 and the Chief Editor of VoiceXML 3.0, I can tell you that the intended audiences were different. VoiceXML 2.0/2.1 was intended primarily to be a standard language for implementing Interactive Voice Response (IVR) applications, and as such assumes an underlying telephony model. There is always a concept of a call in progress and a (spoken or tone) dialogue between a human and a machine. Since the interaction is primarily on the audio channel and must adhere to social norms for spoken interactions, there are very critical timing parameters that are completely baked into the solution, as is appropriate. The HTML Speech effort is about adding speech capabilities to a visual browser, where there may not be an explicit notion of time-sensitive dialog or of audio as the primary (or only) channel. While you could build audio-only dialogs using HTML Speech capabilities (whatever they end up being) and you can build non- dialog, non-audio-exclusive applications with VoiceXML 2.0/2.1 (and even more easily with the new VoiceXML 3.0), neither language is properly suited to do the other's job. It is my personal hope that we will, over the next few years, discover how to transition smoothly from one to the other, so that an app originally written in VoiceXML can, with limitations and a bit of work, be ported to HTML and vice versa. This may become even more appropriate as mobile devices become capable of both and as telephony, audio media, and visual media handling all become truly integrated on the devices. -- dan On Apr 6, 2011, at 12:13 PM, Cesar Castello Branco wrote: > WHAT ABOUT http://www.w3.org/TR/voicexml20/ ? > > IT IS CURRENTLY IMPLEMENTED BY OPERA BROWSER. > > WHY REINVENT WITH A NEWER SPECIFICATION ? > > >
Received on Thursday, 7 April 2011 12:11:36 UTC