This book is an excellent and thoroughly enjoyable reference/tutorial. The book is suitable for use in an advanced undergraduate/first-year graduate class with a prerequisite of one semester of undergraduate E&M. (The authors' preface indicates that this prerequisite is not entirely necessary, but I don't see how you could understand what is going on without it.)
The book can also be used for self-study. In this vein, the book's website contains 1d-, 2d-, and 3d-matlab scripts that are excellent for learning how to actually implement all of this stuff. The third edition weighs in at just over 1000 pages with a price tag of $139, which is $10 cheaper than the 2nd edition was when it came out.
Allen Taflove is, perhaps, the leader in the development and use of this technique. Allen is now at Northwestern. Susan Hagness was a recent PhD student of his (1998) and is now an associate professor in the Electrical and Computer Engineering department at the University of Wisconsin. The authors are at the forefront in the development of applications.
The third edition is significantly larger than the 2nd edition and includes several applications chapters that were cowritten with the major researchers in the field. The extraordinary explosion of application areas for FDTD is captured in the later chapters, and these chapters give students and new researchers a clear flavor of the vitality and interest in the field which extends from the detection of breast cancer to ELF pulses produced by earthquakes. It is refreshing to find authors who so readily give credit to others in their field. Taflove and Hagness have been very gracious in this regard, and as a consequence have a much better book and a very detailed and useful bibliography.
I very heartily recommend this book to anyone who wishes to use FDTD techniques.