Chair's report, September face-to-face meeting in Toronto

The Voice Browser Working Group met on September 20-23 in Toronto hosted 
by Voice Genie.  During the meetings we reviewed several interesting 
systems and software tool kits:

    * AT&T's Florence dialog system [1]
    * RDC--Reusable Dialog Components [2]
    * IBM's reference architecture for SCXML [3]
    * xHMI reusability 

We fine-tuned the CCXML document in preparation for moving from Last 
Call Working Draft to Candidate Recommendation in October.  The Working 
Group also worked on specifying additional assertions for the 
implementation report for CCXML. 

We discussed the Data-Flow-Presentation architecture [4] which separates 
data model from flow model from presentation model. XML and ECMAScript 
are candidates for the data model. CCXML and/or SCXML may be used for 
the flow model. And VoiceXML 2.0, VoiceXML 2.1 and VoiceXML 3.0 may be 
used for the verbal presentation model.  VoiceXML 2.1 can act as a 
verbal presentation "modality" but will need extensions in order to 
synchronize with other multimodal presentation modalities such as 
display (HTML) for multimodal applications

We are considering adding several new elements for VoiceXML 3.0:  

    * New element(s) for speaker verification/authentication--we
      reviewed the VoiceXML Forum's requirements for Speaker
      Verification and Speaker Authentication for VoiceXML 3.0 [5]. A
      subgroup of the Voice Browser working group will develop the
      VoiceXML specifications for speaker verification/authentication,
      while the VoiceXML Forum will review the specifications, provide
      additional use cases, and specify the exchange format for voice
      biometric information.  
    *
    * The <data> element from VoiceXML 2.1--to access data in the flow
      component by the flow and presentation components.  
    *
    * An element similar to <audio>--for presenting media such as video
      in VoiceXML 3.0.

We are also considering enhancements for the <grammar> <audio>, and 
<record> elements and DTMF processing. 

Thanks to Voice Genie for providing an excelling facility for our 
meetings.  The next Face to Face meeting will be in France on Feb 
27-March 3.  

Our revised target publication dates are

    * VoiceXML 2.1--Proposed Recommendation in Mid October
    * Pronunciation Lexicon Specification (PLS)--Last Call Working Draft
      Early October
    * Semantic Interpretation (SIRS)--Candidate Recommendation Early October
    * CCXML--Candidate Recommendation Late October
    * State Chart Extensible Markup Language (SCXML)--Working draft
      Third Quarter 2006

Jim Larson
Co-chair, Voice Browser Working Group

 

[1] http://www.difabbrizio.com/papers/icassp2005-florence.pdf 
<http://lists.w3.org/Archives/Member/w3c-archive/2005Jun/att-0123/Florence_DM_Toolkit_-_W3C_MMI_-_20050627_-_di_fabbrizio_-_rev03.pdf> 


[2] http://jakarta.apache.org/taglibs/doc/rdc-doc/intro.html

[3] http://jakarta.apache.org/commons/sandbox/scxml/

[4] http://lists.w3.org/Archives/Public/www-voice/2005JulSep/0009.html

[5] 
http://www.voicexml.org/resources/SIV_requirements_for_VoiceXML_20050914.pdf