Document:Q9960

I want programming to be as immediate and fluid as drawing and I work with software in a way that minimizes the technical aspects. I often spend a few days creating a core piece of technical code and then months working with it intuitively, modifying it without considering the core algorithms. I use the same code base to create myriad variations as I operate on the fundamental code structure as if it were a drawing - erasing, redrawing, reshaping lines, molding the surface through instinctual actions. In the past year, I have begun removing code from the process of creation. The concept for the work develops entirely through sketches and the final piece is an annotated written description without reference to a computational implementation. The work develops in the vague domain of image and then matures in the more defined structures of natural language before any thought is given to a specific machine implementation. I'm calling this type of program a software structure.
A defining factor in this shift was the work of Sol LeWitt - specifically his wall drawings. A wall drawing is a set of instructions (a text description and optional diagram) outlining a visual structure to be executed on a wall. For example, the program for Wall Drawing #69 from 1971 reads:
Lines not long, not strait, not touching, drawn at random using four colors, uniformly dispersed with maximum density, covering the entire surface of the wall.
LeWitt has written hundreds of wall drawings since their origin in 1968. Each time a wall drawing is reproduced it is different depending on the site and the draftsperson. There is a complete separation of the concept of the work from its perceptual manifestation. The relation between LeWitt and his draftsperson is often compared to the relation between a composer and performer , but I think it's also valid to look at the comparison between a programmer and the entity of execution. LeWitt writes programs for people to execute and interpret rather than for machines. This difference allows him to work in natural language, rather than the limited formal languages of computer code. He is also able to leave ambiguity in his programs, as they will be executed by skilled draftspersons who are able to interpret, rather than a machine which must be told precisely what to do.
This essay explores the potential influences of LeWitt's work on contemporary works of software. I have created three example software structures. To further define the concept of software structures I have written and commissioned twenty-four implementations of these structures to isolate different aspects including interpretation, material, and process.