Re: VoiceXML vs SALT

One might argue, and I will, that VoiceXML is TOO high level. I am not
acquainted with SALT, but a low-level interface is in fact the appropriate
kind of interface for a standard, in my view. Higher level constructs can be
provided as add-ons, freely distributed. VoiceXML does not provide the
flexibility needed to support voice applications of any complexity. In
addition it needlessly (in my view) presents a new programming paradigm (the
FIA) which is flawed in a few respects. Better would be to provide low-level
constructs that can be used in association with rich and well-defined
control languages such as Java or Python. If SALT supports this notion, then
it should be considered to be superior to VoiceXML (as long as the
underlying functionalities of TTS, ASR, telephony, and audio output are
available from SALT).

- Dean Sturtevant (representing the views of myself only)

----- Original Message ----- 
From: "Daniel Burnett" <burnett@nuance.com>
To: "Ildar Gabdulline" <ildar@realeastnetworks.com>; <www-voice@w3.org>
Sent: Thursday, November 06, 2003 8:43 PM
Subject: RE: VoiceXML vs SALT



Dear Ildar,

SALT and VoiceXML can both be used for multimodal applications, and for
simultaneous as well as sequential multimodality.  Just as SALT tags
must be combined with XHTML to form a multimodal language, so must
VoiceXML be coupled with XHTML.  The XHTML + Voice language (X+V), a
submission to the W3C Multimodal Interaction Working Group, provides a
formal means for doing this, although other approaches to using VoiceXML
in multimodal systems have also been demonstrated.  VoiceXML and SALT
can of course both be used in voice-only applications.

In addition to speech markup features, VoiceXML also includes high-level
markup for controlling the voice dialog.  In SALT, the developer has to
worry about many low-level details such as handling errors, managing
dialog flow, and preventing dialogs from getting into the "hanging state"
(See SALT 1.0 section 2.6.5).  These tasks are all automatically taken
care of in VoiceXML.

SALT tightly combines the visual view and the voice modes through its
heavy reliance on ECMAScript.  In contrast, X+V loosely couples the modes.
This enables distributed architectures that place lower demands on
clients and networks and that leverage the burgeoning ecosystem of
VoiceXML servers and applications.

While VoiceXML and SALT are both markup languages for voice and
multimodal applications, only VoiceXML is on a W3C Recommendation track
today, a significant factor in the huge industry uptake of the language.
It is widely deployed and is commonly accepted as the worldwide standard
for voice applications.  It is in use today in high volume, robust, and
reliable commercial systems.  We would be more than happy to provide more
detailed pointers to such systems upon request.


Sincerely,

Technical Council
VoiceXML Forum
http://www.voicexml.org


-----Original Message-----
From: Ildar Gabdulline [mailto:ildar@realeastnetworks.com]
Sent: Sunday, October 26, 2003 7:20 AM
To: www-voice@w3.org
Subject: VoiceXML vs SALT


Hi,

I am relatively new to voice dialogs.
Could you please describe me - what are the differences between VoiceXML and
SALT ?
As I understood for the moment both of them are used for the same purposes -
programming of the dialogs.
If this is correct then it seems that having two standard families is
redundant.

Please clarify the situation, if it is possible.

Thanks,
Ildar

Received on Friday, 7 November 2003 07:40:29 UTC