Distinguishing Visualizations from Pretty Pictures

In 1987, Bruce McCormick led an extensive study of the benefi ts of combining computer graphics and computational-science methods. This culminated in the US National Science Foundation panel report “Visualization in Scientifi c Computing,” in which I was (serendipitously) a participant. Many people identify this event as the birth of scientifi c visualization as a distinct discipline. It was soon followed, for example, by the creation of the IEEE Visualization Conference, thanks largely to Larry Rosenblum, Arie Kaufman, and Gregory Nielson. (A few of us still fondly remember the inaugural 1990 meeting in San Francisco.) The discipline achieved maturity in 1995, when Arie Kaufman became the founding editor in chief of IEEE Transactions on Visualization and Computer Graphics. Today, the worlds of science, computing, and visualization continue to evolve at a dizzying pace. Visual analytics is competing valiantly with visualization per se, and the newcomers big data and cloud computing pose novel challenges to the integration of visualization. While the problems of extracting useful information from ever-growing datasets that dominated the panel’s original concerns continue unabated, here I will ponder the parallel issue of understanding the scientifi c quality of visualization content.