Loquendo believes speech market can take great benefit from speech standards and is continuously supporting their development and deployment. As an active member of the W3C Voice Browser group, Loquendo welcomes the Semantic Interpretation for Speech Recognition recommendation as the last step toward a standard-only speech grammar definition.

Further to the full implementation of the W3C SRGS (voice and DTMF mode, ABNF and XML syntax, signal and text input), Loquendo ASR 7.0 now fully supports SISR specification, enabling the developer to use both "Script" and "String Literal" tag syntaxes, for an easy yet powerful EcmaScript-based Semantic Interpretation.

The support of SISR standard is also available in all Loquendo products exploiting ASR, namely Loquendo Speech Suite, the client server solution based on MRCP, as well as VoxNauta, the VoiceXML/CCXML voice platform solution.

Loquendo implementation was run on an unchanged copy of the official test set http://www.w3.org/TR/2006/CR-semantic-interpretation-20060111/testarchive20051112.zip.

Required features:

• All required features were implemented in conformity with the specification, and the entire test set concerning them obtained a successful result.
• The only slightly difference in Loquendo implementation is global variable's behaviour. These variables, declared into header tags of a grammar, are not read-only (they can be accessed but also modified by the rules of the same grammar). Loquendo agrees and complies with the assertion that global variables must not be shared among different grammars, but thinks it can be useful for developers to share them among rules defined in the same grammar.

Optional features:

• The optional "XML serialization" of the EcmaScript result was implemented and produced expected output.
• The 'score' variable associated to each Rule variable was implemented and successfully tested. The 'starttime' and 'endtime' variables are set to undefined (associating an absolute timestamp to audio data is not always feasible).