AbstractSide effects occur when a computation performs some actions “behind the scene.” Such actions complicate programming activities and the process of maintaining and reasoning about program components. To address these complications, it is first beneficial to expose the side effects as explicit actions whose propagation then can be checked explicitly and enforced by programs and program analysis tools. Additionally, side effects could be confined to specialized sublanguages, which have been designed to deal generically, and in a modular way, with a heterogeneous collection of side effects. Such specialized languages simplify—but do not completely solve—the tricky problems that occur when different side effects interact.
