Understanding the use of diagrams and other components of scholarly systems-related papers may inform us about the papers, the field, and the way we think. We analyze 495 papers containing 1899 figures in a system sciences conference proceedings. We code the diagrams into genres. We find that well-specified diagram types occur less than vaguely specified diagram types. Table types, diagram types, and equations in papers appear to correlate to non-topical aspects of the paper such as the research method used. We relate this idea to current ideas about facets and genre. The work has implications for automated document searching; the practical implication for authors is that more explicit consideration of diagram types might both clarify thinking and make later searching by other researchers easier. The implication for publishers is that collecting component information, and presenting thumbnails, may enhance search.
Jeffrey V. Nickerson