Re: nothing on the spec to clear dtmf buffer ?

I am not familiar with behaviors on that particular platform. 
However, within the specification, it is possible that if you declared a
bargein='false' prompt in the second field, you would clear the dtmf
buffer. I know that's not what you identified as the flow. i.e. block with
bargeable prompt followed by field with bargeable prompt. Alternatively,
you might be able to place an empty block between the prompting block and
the field.


Looking at some voicegenie docs, it looks like there is a platform
specific field attribute 'cleardtmf'. Set it to true to clear DTMF upon
entering. If you are only deploying on that platform, you might be happy
with that option.



Michael Glover
Software Engineer
Interact, Inc.


-----Original Message-----
From: Stephane Boyera <boyera@w3.org>
To: www-voice@w3.org
Date: Thu, 20 Oct 2011 16:11:35 +0200
Subject: nothing on the spec to clear dtmf buffer ?

> Hi
> I've been trying to find a way to clear dtmf buffer before entering a
> form.
> The typical case is very simple:
> 
> *playing a prompt with barge-in authorized
> *then having a form with a menu which plays a prompt (with still 
> barge-in authorized) and capture the input
> 
> what's happening is that because the first prompt is not consuming any 
> dtmf, then any key pressed is passed to the second form and interrupt 
> directly the second prompt and evaluate the key press.
> 
> So in practice, the barge-in in the first prompt skip the second one.
> 
> I've been trying to find a way to clear the dtmf buffer.
> so one (ugly) way to do that is to insert between the two part a fake 
> form that consume all dtmf. that works but this is not really art 
> programming.
> I see also that some platform added an attribute for form
> See e.g. voicegenie and cleardtmf attribute: 
> http://developer.voicegenie.com/voicexml2tagref.php?tag=st_form&display
> =standardtags 
> 
> 
> This does not look very standard compliant. I dind't find anything in 
> the spec for that.
> Did i miss something and is there a declarative way to clear the buffer
> without going through a hack ?
> 
> Cheers
> Steph
> -- 
> Stephane Boyera		stephane@w3.org
> W3C				+33 (0)6 73 84 87 27
> BP 93				fax: +33 (0) 4 92 38 78 22
> F-06902 Sophia Antipolis Cedex,		
> France
>

Received on Thursday, 20 October 2011 20:29:53 UTC