It seems reasonable to assume that published works are related if they are cited together in other works. It further seems reasonable to assume that networks of such relationships could indicate the structure of a subject or discipline. The authors refine this idea, and use it to analyze a large amount of data in their search for representations of “the structure of science.”

This book is made up of nine chapters and five appendices, plus a bibliography (of 151 entries), together with 81 figures and 53 tables. Chapter 1 points out that this approach uses graph theory and strives for simplicity, while chapter 2 argues for the value of a visual presentation of structural information. Some basic notions of graph theory and some background work are presented in chapter 3. Chapter 4 describes some approaches to modeling dense sets of relationships, focusing on the value of a graph representation, and, specifically, extracting an appropriately representative spanning tree from a graph (the simplicity principle), which requires some algorithmic care both in the extraction and in the spatial presentation; these are referred to as scientograms. The Web of Knowledge that was used as a data source is described in chapter 5. Chapter 6 is a detailed presentation of the method used, with options considered and chosen. The results are presented in chapter 7. This chapter contains a large number of figures, many of which are almost unreadable because of the font size needed to fit them on to the printed page, and the discussion refers to colors that are not detectable in the monochrome printed versions; since the figures are such an important part of this entire project, not having usable versions in the book is a significant limitation. Machine-readable versions of the figures are available on a Web site that is referenced in the book. Chapter 8 is a discussion of results, and chapter 9 is an epilogue commenting on the significance of the project. The appendices contain information about sources of data, and some calculated values based on the data.

The authors give a readable explanation of an interesting project. They also present some results representing the international scientific enterprise, as well as some more specific results for Spain, and are able to make some inferences based on the differences in the two sets of data. The work described here is a significant contribution; it will be interesting to a broader audience than the information retrieval research community when more of the analysis and presentation can be automated for more immediate use.