Michael F. Goodchild, National Center for Geographic Information and Analysis, and Department of Geography, University of California, Santa Barbara, CA 93106-4060, USA.
Spatial information that is, information about objects located in a spatial frame such as the Earth's surface has long been recognized as presenting special problems for computing. As early as 1972 the term spatial data handling was being used to refer to the activities of a small but dynamic community of researchers who were committed to exploiting electronic data processing in order to increase productivity in such areas as map compilation and editing; map measurement; and spatial data analysis. Spatial information is rich in high-level structures, and although each of the classical models of database management that began to emerge in the 1960s had something to offer to this application area, neither the relational model nor object-oriented modeling provide a perfect fit. The relational model handles topological relationships well, but lacks the means to represent complex hierarchical relationships that span spatial scales; while object-oriented models handle both topological and hierarchical relationships, but have difficulty dealing with phenomena that are essentially continuous in space.
Books that elucidate the complex story of spatial databases are few and far between, and thus this book is especially welcome. It covers the entire field, from representation through query to analysis, in a style that is clear, logical, and rigorous. Especially welcome is the chapter on data mining, which addresses both traditional spatial data analysis and also new techniques that have been developed in the past few years to take advantage of today's high-speed computing in processes of automated search for anomalies and patterns in very large spatial databases. It is intended for computer science students, as reflected in the sequence of topics and the style of presentation, but will also be useful for students from other disciplines looking for a more rigorous and fundamental approach than is provided by most textbooks on geographic information systems (GIS).
The importance of spatial databases is growing rapidly, partly as a result of the recognition that their applications extend well beyond the traditional domain of GIS. Location and time are powerful ways of identifying and characterizing information, because many data sets have footprints in space and time. This is obviously true of maps and Earth images, but is also true of many reports, books, photographs, and other types of information. Thus location is a powerful basis for search, and for finding relevant information in distributed resources such as the World Wide Web. There is growing recognition that space (and time) provide important ways of integrating information that go far beyond the traditional domains of spatial databases and GIS. The last chapter of the book discusses some of these, and gives a sense of why many believe that the importance of spatial databases is likely to continue to grow rapidly over the coming years.