This paper is named purposely to remind readers of the Dijkstra classic on the trouble with using the “go to” statement in programs. It provides a sound explanation of the problems with choosing the rainbow color map (the colors in the order of the rainbow) for visualization of data. After demonstrating how using the rainbow color map is “confusing,” “obscuring,” and “actively misleading,” Borland and Taylor present data covering the IEEE conference proceedings from 2001 to 2005. This demonstrates its surprising frequency of use. One reason is that the rainbow color map is the default choice in many visualization tools. The paper closes with a discussion of how to make a better map for visualization by explaining the types of data (such as nominal, high-frequency ordinal, interval, and ratio).
The paper does not suggest that writers may choose to use the rainbow map because of an appeal for all of those colors, but this may be a reason. One could question why the rainbow map, since it is an intrinsic aspect of nature, is not better wired into our brains. In contrast, it is easy to accept the benefits of perception in the blackbody radiation spectrum.
The paper has adequate references and the illustrations are well done. While Borland and Taylor may have been constrained for space, some real examples would have enriched the reading experience.