The research started with an investigation into action.
The starting point was ballet choreography and joint rotation, inspired by the work of Tom Calvert at Simon Fraser. I first encountered the work on an NSERC visit, and later through Diane Thompson.
My first explorations into Language-based Action was based on C, and focused only on joint rotation, directionality, and time. Everything ran on a single PC.
The next versions attempted to build a layered representation, where joint rotation and bone movements (the physical reality) was separated from internal information processing. Internal information processing in turn was layered into time coordination and movement control separated from script execution. The different processes ran on different PCs.
An internal geometry separate from a stage geometry was introduced subsequently. Also, the simulation was moved from C programs executing on local PCs and communicating via LAN to the Web with SVG and the use of Active Server Pages.
Current simulation work is focused on the integration of vision into the control of action.
Current theoretical work is focused on exploring the integration of learning.