Plum IVR Support

I am working on integrating our speech application with your platform and have quick question. Your recognizer seems to want to return an array of results for each utterance it recognizes (ala-nbest). Is there a way to disable this?

For example, the following VoiceXML causes the engine/gateway to choke:

<filled>
<assign name="_inputmode" expr="application.lastresult$.inputmode"/>
<assign name="_confidencescore" expr="application.lastresult$.confidence"/>
<submit next="Processor" method="get" namelist="_inputmode _confidencescore" fetchaudio="audio_files/percolate.wav"/>
</filled>

This is what I see in the logs:

Found grammar match
hypothesis #0: number (0.7020)
Saved utterance 1 for DNIS 3140 on channel 2
errmsg TypeError: document.lastresult$ has no properties line 1 linetxt tokentxt
received event: error.semantic.ecmascript

Thanks in advance,

Jose de Castro
Application Architect

Hello,

Could you please provide a more complete example of the form item and IVR grammars you are using. With a very simple example both the application.lastresult$.* and application.lastresult$.* are filled in accordance with the VoiceXML specification:
This simple IVR example does work, however it is possible you are using the IVR platform in some unanticipated way that we will need to resolve. Please provide a simplified but complete IVR code example that demonstrates the behavior so that we can correct the behavior in the IVR platform.

Regards,
Plum Support

Hello,

Actually taking a second look at your IVR code, the IVR platform is complaining that document.lastresult$ does not exist. This is not a predefined variable in the VoiceXML specification and is not supported by the Plum IVR system. You should make use of application.lastresult$ for referencing any previous input on the IVR system.

Regards,
Plum Support