Moody Robots

In one of the founding posts for this blog (Media Art Was the Booby Prize) I described a hierarchy of system behaviors with which to classify and direct my bricollage, see: Taxonomy. Since then I've been around and around on the difference between Responsive and Interactive, and think I may have a hammer to apply to it...

Responsive, not to put too fine a distinction, responds to inputs. My doorbell example is a bit flippant (it's what you've come to expect, yah?), but we might also think of amoeba who shy away from noxious chemicals and most-all Kinect-based Video Art that I've seen.

Interactive changes it's response over time. Interactive systems have internal state that is influenced by external events, and, with luck, those external events are in-turn influenced by the system's responses.

Adaptive is further up the tree. It remembers changes it has made. Ideally changes which somehow improve response. But that's a little beyond me in this current iteration.

A simple way to make an Interactive system is to make it moody. In the case of We Are Experiencing Some Turbulence this could be a child's sliding scale of:

1. Asleep,
2. Bored,
3. Interested,
4. Playful,
5. Excited,
6. Tantrum,
7. Shutdown.

Which is interesting because the expressed states form a loop from shutdown back to asleep. So it could be implemented with a simple wrap around counter that is incremented and decremented based on how intense it's inputs are and how long they last. And, with no input, it can slowly loop through all it's behaviors so it's not just sitting around waiting to be stimulated.

Because things always go better with illustrations, here's one:

The Mood is a function of the Input over Time, and the Output is a function of the Input and the Mood. Inputs may have paradoxical effects when combined with extreme Moods, e.g., high intensity Input during a Tantrum could force a Shutdown of the Output, or low intensity Input in a Bored state might briefly appear to be Excited

Now to get to work...