The star images are an ongoing project comprising thousands of individual photographs originally made as early as 1986. Like many young photographers at their start, I wanted to access the widest possibilities, despite having vastly inadequate means. A picture that resulted from just pointing a camera up at a night sky was inevitably a disappointment: a few blurred dots against an otherwise empty negative.

What for years felt like a disconnection with lived experience eventually provided a way of engaging it obliquely. By foregoing the hope of ever attaining its worthy illustration, the discontinuous aspect of the photographic process became instead its more realistic model.

Beginning in 2005, the archive of photographs from all previous years up to that point was scanned, and then digitally compiled into a single composition. The indication of each star is indexically accurate—each dot does represent some quantity of starlight registered on film or sensor—but the sum total arrangement as depicted has no realistic corollary in any constellation. The black sky background has been removed, and replaced with a plain field of flat and obviously synthetic color.

At the beginning of every year since, all of the new photographs taken during the previous twelve months are added onto the compilation of past years, generating a single new image. These variations usually appear very slight in the near term, as the density of information accumulates at only a minute scale from year to year. Whether that density of stars or the empty space between them ultimately predominates will depend on the duration of the undertaking and the sum of experience it parallels.

A new background color is selected at random for each year by an algorithm that converts numerical patterns generated by the observation of atmospheric noise into one of sixteen million possible tones. (For more on this process, click here to see random.org) The location, timing, and composition of individual photographs within the master compilation are also completely non-systematic, and no records are maintained of their sources.

As printed objects, the images measure 120 inches x 63 inches (304 cm x 160 cm), but have no “correct” orientation—they can be viewed as horizontals or as verticals, and no side is any more accurately the top or bottom than any other. Regardless of orientation, the images are wide enough to completely fill the field of view when standing at a distance sufficient to easily read the definition of individual stars.