Description

This is a rough history of some early experiments in developing a "tracker" class. While I don't consider the applets here as finished works, the collection of them may be interesting as an insight into the creative process that eventually led to something like the Scrawl applet.

Base Class & Rendering Variants

The following four applets show the base class as it progressed through point, ellipse, line and rectangle variations. These are all "mouse-follower" applications.

Tracker01 points Tracker02 ellipses Tracker03 lines Tracker04 rectangles

Following Something Other Than Mouse

At this point a little detour occurred. I wanted the tracker class to operate on input other than the mouse, so spent a little while investigating some simple "Leader" classes that the tracker could follow. (more interesting leader classes can result in entirely different works not yet presented here - todo) These applets are less interactive, though you can still control the tracker parameters.

Leader01 random Leader02 sines Leader03 curves

Assembling Multiple Trackers

At this point the single tracker seemed pretty stable and I began experimenting with combinations of trackers. That is, trackers following other trackers, or multiple trackers following at different rates, et cetera. These applets again are interactive mouse-followers.

Pair01 two trackers first follows mouse second follows first draw line between Quad01 four trackers in two pairs each follows mouse draw line between second of each Chase01 50 trackers first follows mouse rest follow each other at different rates Pack01 50 trackers all follow mouse at different rates

Visually, the output of "pair" and "quad" appear similar. The difference is in how they respond to the mouse. "Quad" is one step further removed from direct mouse response, and is almost like dragging around a spinning top - you sort of need to plan in advance for its inertial response. "Pack" was close enough to the initial goal of these experiments that I'll conclude with it.

Where To Go From Here

The work that appeals to me is finding interesting leaders for these trackers. For instance, have the leader iteratively evaluate a chaotic attractor, or trace the outline of a bitmap, or some other algorithmic course. The additional "looseness" of the trackers may create an interesting interpretation of such courses.

Feel free to borrow/extend any of the classes if you'd like to experiment.