Drawing Lines with Simple Programs: Some Artistic Experiments with NKS

John M. Bacus
Rice University School of Architecture

The history of the creative act in the twentieth century can be framed as a continuous and deliberate abstraction of the artist from the work of art—with an increasing emphasis on the processes of creation over the products. Two particular examples of note in this abstraction are the recognition of autonomous art (as defined by Adorno and the Frankfurt School) and the rise of automatic works like those performed by John Cage. I believe that a centralizing idea for both these threads may be found in the notion of the automaton—an entity both autonomous and automatic which, in a Deleuzian sense, becomes simultaneously the creator and the creation of a work of art.

Connections to NKS in this framing of the arts are quite clear. The notion of an automaton itself is of course central to NKS, but more importantly, the practices of NKS research (performing searches of problem spaces, reliance on visualization, etc.) closely parallels the processes of post-modern artistic creation. In the twentieth century, there were many attempts to marry art and science. I believe that NKS allows for this marriage to work in far deeper and more collaborative ways that it ever has in the past.

Architecture is typically about solids and surfaces, but these present not inconsiderable computational problems. A brief foray into the complexities of computational origami (which might be used to compute constructible and inhabitable form from flat materials) convinced me of the non-trivial problems ahead. To follow NKS into this problem space, I needed to simplify things as much as possible. So I started my research by looking only at lines.

There is a precedent for this in drawing. First year drawing students often begin with simple line drawing—forgetting at the beginning about form, weight, light, etc. in an effort to focus strictly on the most elemental practices of drawing. My initial thought was that I should see if I could use NKS to draw lines as simply as a child.

One particularly useful analog here is found in blind contour drawing, where the draftsman must make a drawing using only one continuous line. As an added obstacle, the draftsman is prohibited from looking at the drawing until his pen is lifted. A computational analog for this kind of drawing is the L-system, where a string of instructions are fed sequentially to a drawing entity which travels across the paper one unit at a time until the instruction set has been exhausted, leaving a single continuous line as a historical record of the path it has followed.

To visualize cellular automata with an L-system, I mapped drawing instructions to digits in the CA. For example, in an elementary two-color CA, a 0 bit was mapped to an LF instruction, and a 1 bit was mapped to a RF instruction. There are several different ways in which this could be done, but even this simple mapping gave quite interesting results.

Created by Mathematica  (April 20, 2004)