Outcomes of Graphical Debugging


eye
The Eye

Principles of a new type of computer art

During the development of computational models of physical systems or mathematical structures it inevitably happens sometimes that the program's behavior differes from expectations, so that the need appears to understand these differences. The graphical representation of computed quantities is much more expressive than printing out diagnostic numbers. If the data to be represented are two-dimensional, it is convenient to convert numerical values to colors and thus represent the data as a color image. In various investigations I created thousands of color images this way and the surprises and the learings were manifold. Sometimes I found such an image surprising and appealing as a picture, as an aesthetical object. Probably much too late, I got the idea to store such surprising results of my program debugging activities in an 'art collection'. In the first acts of collecting I was tempted to intentionally modify the outcome (as the programmer, I was in control of all the parameters which let that particular image appear on screen). For instance, in the eye image above, one could try to center the small circle with respect to the frame. Soon I came to consider such attempts as naive and foolish. If anything was fascinating with these pictures, it was the fact that they were created for no reason else than for understanding the behavior of a (typically flawed) program. No artistic intent whatsoever. The only influence of art-related judgements is in the decision to put a created frame into the 'arts collection' (which implies giving it a title).
Last modification: 2011-10-18