Java applications are often seen as slower than native code apps--it's part of the cost of platform independence and ease of use. However, there are speed-up techniques modern coders ignore with fast CPUs and huge memories to cover for them.

Java Performance Tuning covers techniques applicable to coding in any programming language with those specific to not just Java but particular versions. Shirazi starts with an illuminating discussion of real speed versus the user's perception of speed, threading, caching, streaming, bench-marking and choosing what to measure.

Examples of generally applicable speed-up techniques include obvious examples such as moving unnecessary calculations to the outside of loops while Java-specific techniques include avoiding garbage collection. Some techniques are also applicable with other object-oriented languages; for example, reusing objects instead of recreating them.

The text is clear and readable. Everything discussed is shown as code examples with percentage timings to give an idea of the gains you can make. There is also extensive discussion on choosing algorithms.

Shirazi's coverage of distributed programs is particularly enlightening as they suffer from different bottlenecks, at least from a user point of view. Making a user feel the program is faster is often more a matter of retaining the user's attention than actually increasing speed--the Internet, for example, is unlikely to respond to code enhancements.

Because a better understanding of the way your programs work also makes for greater memory efficiency, more robustness and a better user experience--as well as more speed--this book will dramatically improve your programs. --Steve Patient