W3C home > Mailing lists > Public > www-voice@w3.org > October to December 2003

RE: VoiceXML - limit max call time?

From: Royles, Chris <Chris.Royles@vicorp.com>
Date: Thu, 20 Nov 2003 11:19:29 -0000
Message-ID: <A10DE5600B42CB4EABF446C98C053B8E047BBC@Bulldog.vicorp.co.uk>
To: "'www-voice@w3.org'" <www-voice@w3.org>

Thanks to the members who responded to my question regarding total prompt
timeout. The excellent suggestions indicate one of 2 approaches.

1. Use ECMAScript to query time in order to terminate the prompt. This would
certainly provide conditions on blocks that would only execute if the
particular time related condition is met. However, it is not clear how the
script could throw an event (e.g. noinput) at a particular 'maxtime' within
a dialog. Also constant poling within a script in order to detect a timeout
event is not really appropriate.

2. To ask a VxML vendor for a custom tag extension.

We are currently experimenting with the ECMA solution. The vendor specific
tags are not really an option, we need to find a vendor to implement them,
we are also reluctant to introduce dependancies on custom vendor tags. 
Is there a formal process for raising feature requests for introduction in
any future specification? Can you please advise me on the correct process. 

I have 2 additional issues that I currently have under investigation.

1. The termination of the VxML session and the playing of the prompt buffer.
The specification appears to indicate that <disconnect/> will NOT flush the
prompt buffer, while the <exit/>, <return/> or not having a <goto/> WILL
flush the prompt buffer before exiting the context and transitioning to the
end of the call. Can this observation please be confirmed/clarified.

2. Ensuring under normal conditions, a single cleanup operation is carried
out.
If I wish to maintain a server side session, and have the ability to
invalidate that session at the end of a call, does anybody have a solution,
or recommendations for ensuring that a <submit next=http://..../logout/>
will get through to the server in all standard disconnect/error/exit cases?
(assuming normal conditions and connectivity) I am currently using a root
document that is shared between all my pages, this root document contains an
extensive list of catch statements which then carry out the submit, it is
starting to look a bit scrappy.

With thanks and regards, Chris Royles

Vicorp			Dr Christopher Royles
Wexham Springs		Senior Software Engineer
Framewood Road		+44 (0)1753 660 583
Wexham			+44 (0)1753 660 501
SL3 6PJ			chris.royles@vicorp.com
Great Britain		http://www.vicorp.com

To give you some background. I work for Vicorp, we provide service creation,
design, development, deployment and management tools. We therefore work with
a broad range of voice browser and application server implementations. We
rely heavily on the VxML specification to provide us with our baseline. We
welcome any updates to the specification that improve the predictable
operation of VoiceXML browsers irrespective of implementation.
Received on Thursday, 20 November 2003 06:19:40 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 30 October 2006 12:48:59 GMT