Computer simulation is one of the most active and productive areas of research in molecular and condensed matter physics and materials science. The literature has ballooned in the past several years and great progress in both methodology and important applications has been reported. In Understanding Molecular Simulation, Daan Frenkel and Berend Smit, senior Dutch investigators in the area of molecular simulation, have provided a treasure. The book is a marvellous mix of just enough formalism with an informal and readable style, sufficient detail to understand methodological advances, appropriate mathematics without rigor mortis and penetrating discussions into some of the most important (and, for the nonspecialist, nonobvious) advances in the field.
In their preface, Frenkel and Smit state that "we intend to give unified presentation of those computational tools that are currently used to study the equilibrium properties and, in particular, the phase behavior of molecular and supra-molecular substances." This limitation does hurt a bit: Some of the most important and active areas, such as ab initio molecular dynamics simulations, quantum dynamics simulations and hopping models, are absent (although there is a brief but successful and illustrative discussion of a classical limit of the Car-Parrinello ab initio molecular dynamics).
The book's indexing, layout and readability are good. There are a few disturbing typographical errors, but they will not really deter even a casual reader.
Understanding Molecular Simulation succeeds admirably in its intentions. It focuses on the two standard methodologies of molecular dynamics and Monte Carlo, and, through comparison and contrast, the strengths and shortcomings of each are made clear. This is a useful text: It has what the working scientist needs to know, and it provides extensive illustration, including simple computer algorithms, data, test case analyses, appropriate references and methodological comparisons. It is a unified text in a field in which most of the published work consists of collective conference reports. In everything from notation to use of such important concepts as generalized Lagrangians in the development of new ensemble sampling methods, the unification is of great help.
There are previous books in this area, and this text does not supplant them. The title is "Understanding Molecular Simulation," and. that is where the emphasis really is placed. Details of particular algorithmic approaches and extensive comparison of simulation results to experiment are found in many other useful books, including those by W. G. Hoover (Computational Physics, Elsevier, 1991), K. Binder, (The Monte-Carlo Method in Condensed Matter Physics, Springer, 1992), M. H. Kales and P. Whitlock, (Monte Carlo Methods, Wiley, 1986), and M. D. Allen and D. J. Tildesley (Computer Simulation of Liquids, Clarendon, Oxford, 1987). Still, this book is more recent than those, and it brings in important recent advances-from variable time-step methods to the Gibbs ensemble approach to phase equilibria. It is intended to be useful, and I think that it is. Much of the complex but necessary technical material is relegated to nine appendices, on such topics as statistical errors, long-range interactions and the Gibbs ensemble.
Frenkel and Smit's book should be useful for scientists, ranging from those beginning graduate school to working professionals. Much is presented, and presented with clarity and perspective. In sum, this book is a good one.