- SRGS is almost a W3C Recommendation
- Describes the expected utterances with a simple grammar
- can be either in Augmented BNF or XML syntax
- Speech processors know what they can expect in a context, hence no need for a very general speech recognition
- this makes speech input feasible
- the “output” is the set of literals described by the grammar
- Speech processors know what they can expect in a context,
- hence no need for a very general speech recognition
- this makes speech input feasible
- the “output” is the set of literals described by the grammar
- The grammar can also be used for non-speech input
- eg, to interpret handwritten input