Can a machine produce every possible image? What are the limits of this kind of automation? Is it possible to practice image making by exploring all of image-space using a computer rather than by recording from the world around us? What does it mean that one may discover visual imagery so detached from 'nature'?

Every Icon progresses by counting. Starting with an image where every grid element is white, the software displays combinations of black and white elements, proceeding toward an image where every element is black. In contrast to presenting a single image as an intentional sign, Every Icon presents all possibilities.