American Scientist

Number by Colors: A Guide to Using Color to Understand Technical Data. Brand Fortner and Theodore E. Meyer. 349 pp. TELOS, (Springer-Verlag), 1997. $44.95.

Brand Fortner is well known for his long involvement with such data-visualization programs as "Plot," "Transform" and "Slicer." Hence it seemed reasonable to expect this book to be filled with really insightful suggestions for and examples of computer-based data visualization and analysis. Surprise! The first part of the book is devoted to a review of human color vision, how video screens and hardcopy printers reproduce color, how different color spaces work, and other fascinating topics. Although presented in a well-structured way that is interesting to read, Fortner's information is not as technical as some folks might need if they actually planned to use it. Pratt's Digital Image Processing (Wiley, 1978, 1991) or Foley and Van Dam's Fundamentals of Interactive Computer Graphics (Addison-Wesley, 1982) include far more quantitative detail as well as equations for converting between color spaces.

The second half of the book discusses data and file formats (covered at greater length in the author's previous book on data structures), and it is only in the final portions of the book that the topic of creating graphics and visualizations comes up. Unfortunately, this is the weakest section of the book. Beyond a few rather trivial "hints" (such as "blue is best used for backgrounds") there is not much real meat. Some of the cautions omit or gloss over important technical details. For instance, the note to be careful about putting colors next to each other, because of difficulties in printing hard copy, skips over the topic of trapping, saying that it is specialized and professional printers know about it. That is not much help for people doing their own printing with a laser printer. A comment that using hue, saturation and intensity to present multiple dimensions of information might be interesting but is seldom done is untrue—I have seen it used quite often and very effectively in various surface-microscopy applications.

Most important, there really are not the substantive guidelines to visualization of complex multidimensional data sets that Fortner's company sells software to handle. More examples of really good (and really bad) presentations would have been very useful. Someone needs to write a version of Edward Tufte's 1983 book, The Visual Display of Quantitative Information (still the most exciting book on examining and presenting data I know), that makes use of the possibilities of computers. I really wanted to like this book. The author has a very readable style, and the information that is there all seems to be quite accurate. I enjoyed reading it right up to the end, but the book's end came before it reached the material I most wanted to read. I just wish that the book delivered on its promise and covered a topic that is becoming quite important to people as the universal use of computers generates larger and larger data sets and promises to help us make sense of them and communicate results. Perhaps the next volume in the promised set of three from this author will deliver that.—John Russ, Materials Science, North Carolina State University