I liked to see nomographs (nomograms) when they were still being used, but I did not always gain a complete understanding of the relationships among the variables from the nomograms. There is a nice orderly contour-line type aesthetic pattern to some nomograms. But the equation behind the nomogram usually gives a better theoretical sense of what was going on; that is, Y varies as log X, taken at various levels of Z, for example. Still, to see the systematic contours--the shape of the equations--is quite helpful.
A great virtue of nomograms is that they are usually multivariate, showing relationships among variables in quite complex systems.
It is surely helpful to have both an analysis of the underlying equation along with nomogram visualization of the curves generated by the equation. Nomograms show how equations perform.
Nomograms remain useful for understanding; their computational use has passed. Computational power is so cheap now, we don't need look-up tables or nomograms; we can just plug the numbers into the equations and solve.
Recently nomograms appear to be used mainly as practical workaday graphics in local applications and particular trades. Nomograms don't appear often enough in scientific textbooks. Nomograms are routinely used in engineering statistics books in order to show operating characteristic curves. The original graphic of Moore's Law is a nice nomogram.
The article by Thomas L. Hankins, "Blood, Dirt, and Nomograms: A Particular History of Graphs," Isis, 90 (1999), 50-80 is excellent, a helpful history and depiction of nomograms. It can be seen at http://www.journals.uchicago.edu/Isis/journal/demo/v000n000/000000/000000.text.html . Hankins' article at this link shows spectacular high-resolution images of many great nomograms.
-- Edward Tufte
