Summary of the Multimodal Interaction Working Group Face to Face Meeting, June 14-15, Somerset, NJ, USA

On June 14th and 15th, 2010, the W3C Multimodal Interaction Working Group[1]
held a face-to-face meeting, hosted by Openstream[2] in Somerset, New
Jersey, USA. 

We primarily discussed the Multimodal Architecture, EMMA, and issues around
the integration of HTML5 and speech. We also had updates on InkML and
EmotionML and several demonstrations of applications developed using the MMI
Architecture.

1. Multimodal Architecture[3]:
	a. Interoperability:
	There is currently an active subgroup working on interoperability
testing of MMI Modality Components and Interaction Managers. We are
investigating what would be involved in setting up a web page to support
interoperability testing. This would allow companies to post information
about their modality components and how they might be used in interoperable
MMI Architecture-based systems.
	b. Issues: 
	We reviewed eighteen open issues for the MMI Architecture, closing
some and scheduling discussions for the rest.
	c. demonstrations: Three demonstrations of MMI Architecture
implementations by Working Group members were presented (see section 6 below
for details.)
	d. We are finalizing the next Architecture Working Draft and expect
to publish it in July. 
	
2. EMMA [4]:
	The EMMA (Extensible MultiModal Annotation) subgroup is currently
discussing use cases[5] for a possible EMMA 2.0 specification. At the face
to face we reviewed some suggestions from Microsoft based on their
implementation experience, and recorded issues for discussion during EMMA
calls. 

3. HTML5 and speech: 
	We believe there is significant industry interest in providing a
markup that enables speech to be used as an input in HTML5 and discussed
ways in which the relevant Working Groups could collaborate. One example
would be by launching an Incubator Group[6].
 
4. InkML[7]: 
	InkML  is currently in Last Call, the comment period having closed
on June 17. A number of comments were received based on the Last Call (see
the public MMI List [8] for comments). The InkML subgroup is working on
responses to the Last Call comments and preparing an Implementation Report
Plan in preparation for moving InkML to Candidate Recommendation. 

5. EmotionML[9]: 
	The Emotion subgroup is preparing a new Working Draft of EmotionML,
with the goal of publishing it in July 2010. The Last Call Working Draft is
currently expected to be published in September, 2010.

6. Demonstrations:  Several demonstrations of applications that use the MMI
Architecture were presented. 
	a. Deutsche Telekom demonstrated a browser-based MMI Architecture
implementation. It includes an SCXML interpreter running within a web
browser and a distributed VoiceXML-based voice modality component.The second
demo of Deutsche Telekom showed a browser-independent ECMAScript-based EMMA
integrator, capable of integrating EMMA messages from GUI and voice modality
components.
	b. Openstream demonstrated its Cue-me multimodal platform demos on
Android and iPad devices with speech, ink, visual modalities  and image
capture, using a  device-based Interaction Manager implemented in SCXML.
	c. Deborah Dahl demonstrated a desktop application that uses the MMI
Architecture to access Windows Speech Recognition engines in six languages
and provides EMMA results. 

Many thanks to Openstream for hosting the meeting and for providing an
excellent meeting environment.

Best regards,

Deborah Dahl
Chair, W3C Multimodal Interaction Working Group

[1] Working Group: http://www.w3.org/2002/mmi/
[2] Openstream: http://openstream.com/
[3] MMI Architecture: http://www.w3.org/TR/mmi-arch/
[4] EMMA: http://www.w3.org/TR/emma/ 
[5] EMMA 2.0 Use Cases: http://www.w3.org/TR/emma-usecases/
[6] Incubator Groups: http://www.w3.org/2005/Incubator/
[7] InkML:  http://www.w3.org/TR/InkML/
[8] MMI List: http://lists.w3.org/Archives/Public/www-multimodal/
[9] EmotionML: http://www.w3.org/TR/emotionml/

Received on Thursday, 8 July 2010 14:13:39 UTC