Cybernetic Lifestyles

stick models view animation

In 1976 I was hired by Nicholas Negroponte at his Architecture Machine Group (predecessor to the MIT Media Lab) to lead a research team to develop software for a recently developed and exciting piece of hardware: a color display system, called a "raster-scan frame-buffer." Computer displays were common in that era but comprised "vector graphics " displays—where the display hardware would move the electron beam to inscribe each line of the correct length and orientation.

"Raster-scan" meant that every line of the display was scanned across, left to right, one-by-one, from top to bottom of the screen. Common in televisions from their inception, raster-scan was a relatively new concept for computer-driven displays. In fact, it was the third such system of its type ever built; Ken Knowlton of Bell Labs and Dick Shoup of Xerox PARC had built the first and second, respectively. Raster-scan was rare because, to achieve the result of scanning every line, the hardware needed to store the color value of every spot of every line—"pixels", a term that was just coming into use. Hence the term "frame-buffer" because the entire frame was stored at any given moment and drove the output of the display.

For the hardware to do this, a considerable amount of RAM was needed—a full megabyte, in practice, to achieve a 512x512 pixel display with 8 bits of color information per pixel. Andrew Lippman and others were involved in the hardware design. The lab was criticized for squandering $14,000 on the 1MByte of RAM needed (I realize you don't need a full me. Of course the price of RAM fell quickly and the lab was vidicated in builing this device, because of the advantages of full color, stable, video-like images, and more. Every computer of today uses a display in the lineage of these early raster-scan devices.

Having done extensive work on vector machines on the PDP-7, I was not anxious to simply abandon what I had found to be an expressive medium, even with its simple lines—or maybe especially because of its crisp, simple line drawing. Moreover, the limits of processing power meant that even even uncomplicated images took a long time to compute. So any attempt at motion graphics required a batch process to write the frames to hard disk. Afterwards each frame was replayed in smooth video (well, mostly smooth; sometimes a disk access caused a pause in the display rate, something apparent in the movies offered here). In the process of creating an animation scripting language that was visual (2-D), I developed a series of increasingly complex simulations that became a short film called Stick Models. [4 minutes silent quicktime] —Paul Pangaro, August 2006