W3C home > Mailing lists > Public > www-voice@w3.org > October to December 2002

missing VoiceXML functionality

From: Jim Larson <jim@larson-tech.com>
Date: Wed, 06 Nov 2002 14:29:40 -0800
Message-ID: <3DC997D4.8060609@larson-tech.com>
To: www-voice@w3.org
Cc: emmett@ejtalk.com

Thanks for your comments.

The best way to process your message is to e-mail it to our e-mail 
archieve.  Right now we have closed all updates to VoiceXML 2.0.  By 
placing it in our archieve, it will be considered when we define 
requirements for the follow on for VoiceXML 2.0.


-Jim Larson


-----Original Message-----
From: Emmett (ejTalk) [mailto:emmett@ejtalk.com]
Sent: Wednesday, November 06, 2002 10:04 AM
To: Jim A. Larson
Cc: Deborah Dahl
Subject: we met at SpeechTEK

Hello Jim,


Nice to meet you at SpeechTEK (at the MSTA booth).


I wanted to remind you of the missing functionality in VoiceXML that we 
talked (I ranted?) about.




VoiceXML does not allow for a simultaneous record AND reco event. This 
is one of the MOST important kinds of information for the developer to 
collect. Every book, article and seminar about tuning and usability 
drones on about how hopeless it is to do spiral development without it.



1. Of course this can be done outside the VoiceXML interpreter if you 
own the platform and can write to the specific ASR system. Not possible 
for mere users of a vsp.

2. The target audience of small implementers can't compete with the vsp.

3. Very powerful, intuitive and (partially) automated tools could be 
used by the smallest of developers to do serious tuning of applications. 
(ejTalk has a tool-centric point of view)

4. It may put a burden on the vsp to slog all that audio around, so it 
would be reasonable to have a multi-tier rate structure. Charging a 
little more when the "save_audio" feature is on would be understandable.

5. This feature may even spur a new sub-industry. There would be enough 
demand for utterance transcription (the missing piece that MUST be done 
by a human -- or better yet a linguist) that audio could be routed 
through a bureau that would add the transcription/annotation that would 
then slide directly into a grammar/dialog tool and provide great 
potential for refining the user experience. (I have many ideas about 
such tools.)




If there are any folks that want to talk further about this then don' t 
hesitate to pass this note and my name along. Or, if there is room in 
the organization for another person full of ideas then I'd like to help.


Hope to talk again soon,





Emmett J. Coin

Industrial Poet, ej Talk www.ejTalk.com <http://www.ejtalk.com/>

"The Allure and Power of Talking with the Machine"

Received on Wednesday, 6 November 2002 17:32:47 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:03:47 UTC