I just dedicated half an hour of recreational hacking to a text2speech
RTT component . It has a single "speak" string input port, and if
you write to it, guess what happens?
The motivation for this came from the observation that while executing
a complex coordination FSM, it is hard to track both the progression
of the FSM (on the screen) and the actual robot motion at the same
time. So by connecting the current state port of the rFSM component to
the text2speech component, I'm hoping to be able to track the robot
motion and the FSM without taking my eyes of the robot.