What is the difference between concurrency and parallelism?

Question

What is the difference between concurrency and parallelism?

Examples are appreciated.

Answer 1

(I'm quite surprised such a fundamental question is not resolved correctly and neatly for years...)

In short, both concurrency and parallelism are properties of computing.

As of the difference, here is the explanation from Robert Harper:

The first thing to understand is parallelism has nothing to do with concurrency. Concurrency is concerned with nondeterministic composition of programs (or their components). Parallelism is concerned with asymptotic efficiency of programs with deterministic behavior. Concurrency is all about managing the unmanageable: events arrive for reasons beyond our control, and we must respond to them. A user clicks a mouse, the window manager must respond, even though the display is demanding attention. Such situations are inherently nondeterministic, but we also employ pro forma nondeterminism in a deterministic setting by pretending that components signal events in an arbitrary order, and that we must respond to them as they arise. Nondeterministic composition is a powerful program structuring idea. Parallelism, on the other hand, is all about dependencies among the subcomputations of a deterministic computation. The result is not in doubt, but there are many means of achieving it, some more efficient than others. We wish to exploit those opportunities to our advantage.

They can be sorts of orthogonal properties in programs. Read this blog post for additional illustrations. And this one discussed slightly more on difference about components in programming, like threads.

Note that threading or multitasking are all implementations of computing serving more concrete purposes. They can be related to parallelism and concurrency, but not in an essential way. Thus they are hardly good entries to start the explanation.

One more highlight: (physical) "time" has almost nothing to do with the properties discussed here. Time is just a way of implementation of the measurement to show the significance of the properties, but far from the essence. Think twice the role of "time" in time complexity - which is more or less similar, even the measurement is often more significant in that case.