I just noticed that the good folks over at TellMe (a relatively new bit of Microsoft) have shipped a beta of their Voice Studio voice application dev tool, based on DSL Tools. Very cool indeed! (I seem to be saying this a lot lately)
I love the process of developing voice apps - One of the last things I did in Microsoft before joining Visual Studio was to build a voice-enabled kiosk for helping folks access local government services. It was seriously rewarding when you managed to have testers use a natural seeming set of sentences to get something done. There's just a certain Star Trek magic to the whole process.
I think part of me may end up getting a similar buzz from extracting structured data from textual DSLs - MGrammar crossed with text extraction maybe - be interesting to go the same route as voice and have a confidence match for the textual DSL and then have a process of refinement from relatively vague textual musings to specific textual models.
Many congrats to the TellMe team (or as I'm steadily learning the colloquialisms over here, GO TellMe!)