Writing concurrent code
Concurrent programming is hard to get right. Even with contrived example applications, the bulk of complexity comes from concurrent code. We obviously want our code to be readable while keeping the benefits of concurrency. We want to get the most out of each CPU on the system. We only want to compute what we need, when we need it. We don't want spaghetti code that joins together several asynchronous operations. Focusing on all these aspects of concurrent programming while developing applications detracts from what we should really be focusing on—the features that give our application value.
In this section, we'll look at the approaches that we might use to insulate the rest of our application from tricky concurrency bits. This generally means making concurrency the default mode—even when there's no real concurrency happening under the hood. In the end, we don't want our code to contain 90% concurrency acrobatics and 10% functionality.
