The Art of Concurrency

Chapter 4. Eight Simple Rules for Designing Multithreaded Applications

Since it is right there in the title of this book, the following sentence shouldn’t come as any surprise: Concurrent programming is still more art than science. This chapter gives eight simple rules that you can add to your toolkit of threading design methods. I’ve tried to organize the rules in a semichronological way, but there’s no hard and fast order to the rules. It’s like being confronted with, “No running by the pool,” and, “No diving in the shallow end.” Both good ideas, but not diving can come before not running or vice versa.

By following these rules, you will have more success in writing the best and most efficient threaded implementation of your applications. You may recognize some of these, since I’ve mentioned a few of them in previous chapters. In upcoming chapters, when discussing the design and implementation of specific algorithms, I’ll try to drop in a relevant reference to one or more of these eight rules to show that they’re not just here to fill out an extra chapter.

Rule 1: Identify Truly Independent Computations

I’ve already covered this first rule seven ways to Sunday, but since it’s the crux of the whole matter, it bears repeating at least one more time. You can’t execute anything concurrently unless the operations that would be executed can be run independently of each other. I can easily ...