Currently only atomic commands can be created at the subsystem level. Being able to define more complex behavior using only
the commands on the subsystem is helpful. For example, a "Drive Straight" command and a "Stop" command could be tied into a sequential command group that drives for a certain distance or time and then stops.
Parallel command groups don't need to be supported because only one command for the subsystem can run at a time. Any parallel group built solely from commands for a single subsystem would be invalid