Concurrency Bugs Explained: Why Shared State Needs Synchronization

Concurrency within a single process involves multiple execution units, such as threads or goroutines, running simultaneously and potentially accessing the same data. A classic example shows how two threads each incrementing a shared counter can produce an incorrect result of 1 instead of 2, because both threads read the same stale value before either writes back. This race condition arises from the three-step nature of a simple increment operation — read, compute, and write — which can interleave unpredictably. Languages like Java, Go, and Python address this through synchronization primitives such as locks and mutexes, which enforce exclusive access to shared state during critical operations. An alternative approach, message passing, avoids shared state altogether by having execution units coordinate through exchanged messages rather than direct memory access.