jakarta-taglibs-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stu Robertson <srobert...@nvisia.com>
Subject Re: RDC: Plan B for getting Nuance to work with RDCs
Date Tue, 14 Jun 2005 17:57:10 GMT
Hi Daniel.  It is confusing to me also :-)

In short, we're unable to get the interpretation to be set to a slot  
by the digits or alpha grammar (or any other grammar actually).  Here  
are some things I've tried:

<?xml version="1.0"?>
<grammar xmlns="http://www.w3.org/2001/06/grammar"
    xmlns:nuance="http://voicexml.nuance.com/grammar"
    version="1.0" mode="voice" xml:lang="en-US"
    tag-format="Nuance" root="numbers">   <!-- I changed the  
attributes to match Nuance docs -->

<rule id="numbers" scope="public">
     <item repeat="1-50">
         <item><ruleref uri="#digit"/>
         <tag>assign(a strcat($a strcat(" " $return)))</tag>
       </item>
     </item>
     <tag><![CDATA[<compatibility $a>]]></tag>
</rule>

This now compiles, but nothings shows up in the interpretation as a  
property.  I've also tried using an unamed slot like <=$a> with the  
same problem.

The only thing that's been close is

<tag> compatibility=$a </tag>

Which creates a property in the interpretation correctly, but sets  
its value to "null".

In all of the above cases the utterance is correctly recognized.  But  
the nBest.js isn't able to get at the semantic interpretation.



On Jun 14, 2005, at 10:39 AM, Daniel Spangler wrote:

> I've kind of gotten lost in this long thread. Stu, could you re- 
> post the
> exact problem that you are currently stuck on?
>
>
>
> ________________________________
>
> From: Stuart D. Robertson II [mailto:srobertson@nvisia.com]
> Sent: Tuesday, June 14, 2005 11:31 AM
> To: Tag Libraries Developers List
> Subject: Re: RDC: Plan B for getting Nuance to work with RDCs
>
>
>
>
>> None of you have any access to Nuance support? Their grammar
>>
> guide sure
>
>> shows CDATA's and assign's and sub-rules and everything that is
>>
> needed in
>
>> GrXML for us [
>>
> http://community.voxeo.com/vxml/docs/nuance20/grammar.pdf
>
>> ]. There is a sample retail application on Page 181 that uses
>>
> GrXML. Have
>
>> you tried changing tag format to "Nuance" i.e. try using the
>>
> same grammar
>
>> root element as the sample shows? I guess you have tried
>>
> everything by
>
>> now.
>>
>
> Hmm, I didn't notice that before, and did not try it.  It makes
> me really wish I were home so I could try it out now (I can only
> test at home b/c of the firewall at work).
>
> Like you, I'd much prefer getting SRGS grammars working than
> have to make a big switch.  Ideally, we could just apply some
> xslt to the main grammars (those based on the final
> recommendation version of the standard, not the one from 2001
> that Nuance uses).  I'll test this as soon as I can and post
> results.  I'll also see if I can get their example working as
> if.  If that works, you're right...what we're trying to do isn't
> really different.
>
> Glad you like the wiki page :-).  I certainly would have found
> that one useful, so I suspect others will too.
>
> Stu
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: taglibs-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: taglibs-dev-help@jakarta.apache.org
>
>
>
> ________________________________
>
> avast! Antivirus <http://www.avast.com> : Inbound message clean.
>
> Virus Database (VPS): 0524-1, 06/14/2005
> Tested on: 6/14/2005 11:30:06 AM
> avast! - copyright (c) 2000-2004 ALWIL Software.
>
>
>
> ______________________________________________________________________ 
> __
> This email has been scanned for all viruses by Netscalibur Mail  
> Scanner, powered by MessageLabs. For more information on a  
> proactive email security service working around the clock, around  
> the globe, visit
> http://www.netscalibur.co.uk/scanner/index.html
> ______________________________________________________________________ 
> __


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message