Review by Choice Review
Shampine, Allen, and Pruess have written a fairly traditional introduction to numerical computing for upper-division undergraduates and graduate students, but with the new twist popular in recent years that the source code (FORTRAN, C, C++, and Matlab) for its subroutines and programs is available using the Internet World Wide Web. These programs are integrated into the presentation at several places in the text. Because code is available electronically, no pseudocode is included. The authors also use case studies to illustrate more advanced material; although these are generally interesting, they often contain terms/concepts with which some readers may not be familiar. Indeed, the authors are sometimes quite optimistic about their readers' backgrounds (e.g., the Galerkin method discussion in Case Study 2). The book begins with material on computer arithmetic and floating point errors; it next discusses basic numerical methods for linear systems, polynomial interpolation (including splines) and root-finding methods (bisection, Newton's method, secant method); the final chapters deal with basic and adaptive numerical quadrature, and basic methods for solving ordinary differential equations (including Runge-Kutta). The style of presentation is clear and precise, but in places there seem to be a number of misprints. The misprints are not serious enough to mislead the informed reader, but they may be more troublesome for beginners. A good basic reference. J. D. Fehribach Worcester Polytechnic Institute
Copyright American Library Association, used with permission.
Review by Choice Review
