Currently there is a "settings" XML file in "my documents/mayhem" that stores the events, the ability to have multiple XML files accessed by seperate instances of mayhem would create huge opportunities for mayhem. Storing the settings file in the mayhem folder would suffice (then multiple folders could be created, or at least reference the XML file in the core.dll or such so the location the file is stored could be changed manually for every instance. Having both speech recognition and text to speech could create a very crude Artificial Intelligence to control things like lights, open/close programs, lock/unlock doors, and check weather all through voice commands, But a having a way to enable/disable the entire AI library would be a must.

Parent Instance ( CMDs to launch/terminate any Sub-Instances such as below )

          Mayhem-AI ( has all voice cmds stored in separate XML from Parent Instance for Artificial Intelligence)

          Mayhem-Media ( Could house all media CMDs include PhoneRemote CMDS )

          Mayhem-Remote (Open/Close garage door, lock/unlock any house door, Enable /Disable alarm system ) 

The benefit of having a sub-instances is being able to turn on/off entire sets of cmds with one key press or voice cmd.

Edited:  removed a bad idea from post that would have created to complicated of program for intended audience.

I like these ideas, and I totally agree, Mayhem easily enables anyone to use speech recognition, one of the most sought after ways to control their computer (and just comes off as totally awesome).

We definitely want to enable people to save and load configuration files, but we believe there are more user friendly ways to group connections than having to load files. It's something down the road we will definitely look into.