Action Semantics - Semantic Entities

Semantic Entities

An important part of action semantics that gives it a modularity not seen in previous programming language semantics is the use of first-order semantic entities. First-order refers to how, unlike in denotational semantics, where a semantic function can be applied to another semantic function, in action semantics, a semantic entity cannot be applied to another semantic entity of its kind. Furthermore, the semantic entities utilized by action semantics broaden the framework’s ability to describe a programming language’s constructs by serving to denote both program behavior that is independent of any particular implementation and the way in which parts of a program influence the overall performance of the whole. The appropriately-named action notation is employed to express the three types of semantic entities found in action semantics: actions, data, and yielders. The central semantic entity in this framework is actions, with data and yielders occupying supplementary roles. More specifically, actions are the mechanisms through which yielders and data are processed. An action, which can occur within another action, is a step-by-step representation of program behavior, where each step accesses current information, changes current information, or does both. Yielders appear within actions and only access current information. A yielder entity is capable of being evaluated, and when it is, the product is a datum entity.