|
|
 |
May 2004
Potentially Breaking Changes
The most important changes are listed towards the top of the
table, with minor ones later. Some of the more complicated changes have
links to the appropriate section of the VoiceXML 2.0 draft (e.g. §5.2)
for more detailed information.
| Relevance |
Description |
Recording
|
When recording using the <record> tag with the default
enctype value application/x-www-form-urlencoded, the binary data posted by the browser
is encoded with iso-8859-1 encoding. Previously the binary data was posting using the Windows
default encoding. This modification should result in better audio quality for the saved audio
files on the backend server. Backend code that reads the binary data should use the iso-8859-1
encoding.
|
New Features/Enhancements
| Relevance |
Description |
Spanish Language Support
|
The browser now supports Spanish language. See the
BeVocal Spanish Language Support Summary for more details.
Applications now can have Spanish language grammars and perform Spanish language recognition.
|
Grammar Compiler Tool
|
The Grammar Compiler tool now supports Spanish language grammar compilation. See the documentation for
Grammar Compiler Tool for more details.
|
Audio File Recognition
|
The browser now supports recognition from an audio file. This feature is helpful in
testing and tuning applications which perform recognition against prerecorded audio files.
See the documentation for <field>
tag for more details.
|
Outbound Calls Information
|
The Log Browser tool allows you to view two additional fields,
"Outbound RequestID" and "Outbound Disconnect Reason", using the Customize View button.
Only outbound calls have useful information for these fields.
|
Voices
|
There are two additional voices available: Vocalizer Female Spanish and RealSpeak Female German. Refer to the
VoiceXML Programmer's Guide for complete details.
|
Log Browser
|
The
Log Browser tool can now search logs by a single Session ID.
|
Vocal Player
|
The
Vocal Player tool can now search logs by a single Session ID.
|
VoiceXML Checker
|
The VXML Checker
tool has a new "Activate" button.
You can activate the current application from within the VoiceXML Checker without going back to
the File Management page.
|
File Management
|
Audio files uploaded using File Management tool can now be played by
clicking on the name of the file.
|
Enrollment
|
During the enrollment process, if a spoken phrase clashes with an
existing entry in the enrolled grammar, you can get the phrase id with which it clashed in the new
shadow variable clashedPhraseIds. See the documentation for
<bevocal:enroll>
for details.
|
Audio
|
The new property,
bevocal.audio.outputvolume, allows you to adjust the output volume from the platform.
|
Spanish Language Support
|
There is now Spanish language support for all standard built-in grammars.
|
Spanish Language Support
|
The Grammar Compiler Service and the Enrolled Grammar Removal Service now support
Spanish language with the additional parameter, xml:lang.
|
Known Bugs and Limitations
| Bug |
Description |
5144
|
DTMF recognitions produce
incorrect values in the utterance property of the shadow variable and the application.lastresult$ variable.
Currently, utterance is set to a string of the form DTMF-1
DTMF-2 DTMF-pound. The VoiceXML 2.0 specification requires that the
value be set to a simple string of digits and symbols such as "12#".
The specification also requires that the termchar be removed from the
end of the utterance string, which currently is not done.
|
6305
|
The numeric scale of the
confidence property of shadow variables and application.lastresult$
now
correspond with the VoiceXML confidencelevel
property. The confidencelevel property uses an exponential
scale to map the VoiceXML property values 0.0-1.0 to the Nuance values
1-100. However, the confidence properties of shadow
variables and application.lastresult$ use a linear mapping.
|
Bug Fixes
| Bug |
Description |
8136
|
Hotword recognition in Vocal Scripter no longer throws an error.internal.
|
7178
|
When a 500 Server error occurs during a fetch, the content of the page,
if any, is printed to the VoiceXML log file.
|
8286
|
With the 2.4.2 release, JavaScript execution is controlled by time
of execution and stack depth limit. The stack depth limit has now increased to 50 from the old value
of 25.
|
8439
|
VocalDebugger tool now works for VocalScripter calls.
Previously it worked only for telephony calls.
|
8682
|
When using builtin street grammar, the street name result would have a consistent value now. Previously it had either mathilda avenue or mathilda_avenue based on user response. Now if the street name has multiple words, the result would have spaces in between them.
|
|
