Over the years, I've seen and been impressed by a number of domain-specific modeling languages for audio processing. This YouTube video of Reactable simply blew me away though (tip of the hat to Risto Pohjonen). As my earlier "Creating symbols" post shows, I'm a fan of direct manipulation user interfaces, and Reactable takes that to the nth degree. I imagine integrating with version control could be an interesting problem, though...

Perhaps the best example in the Digital Sound and Music category is Kyma, the sound synthesis system of Symbolic Sound. Kyma was used on Star Wars Episode I by Ben Burtt, the "father of sound design". It was also used to make the new THX logo sound shown before all THX films. It has since been used for Master and Commander and Finding Nemo, as well as Attack of the Clones.

Whilst many parts of Kyma could be regarded as DSM, the clearest is the Signal Flow Editor (half-way down the page). The essential elements of the language look the same as in 1991, as can be seen on p76 (PDF p8) of this article by Kyma's designer, Carla Scaletti:

• Computer Music Languages, Kyma, and the Future (PDF) Computer Music Journal v26n4: Language Designers on the Future of Computer Music - Winter 2002

The article starts with an interesting section on what makes a (domain-specific) language successful.

If you want more direct manipulation, take a look at the videos of Osculator, which lets you control Kyma with a Wiimote -- the light sabre is a must-see!