Base-type languages: Program design by upper semantic approximation

Abstract

The mathematical model of programming languages considered in this paper has a relatively simple and regular form: the program is represented as a collection of tree structures supported by one another. This model allows highly efficient verification of its formalized properties and admits most features of high-level languages. Its limitations include unavailability of recursion and static definition of constructs. The proposed approach to program design involves a sequence of program modifications: after each modification the set of all possible behaviors of any operator becomes a subset of the set of all possible behaviors of the same operators in the previous version of the program. From the point of view of changes in program form, modification involves adding new hierarchical structures (trees) without altering the existing structures. An essential aspect of this approach to program design is its ability to detect absence of deadlocks. Section 1 provides a description of the structure of the languages. The main distinctive features of our model are the following: recursion is not available, so that program schemas are specified as a special representation of partial orders; the definition of denotational semantics explicitly reflects structural properties through the events {open_quotes}start complex{close_quotes} and {open_quotes}end complex{close_quotes}; a unifiedmore » technique is available for description of algorithm structures and data structures. Despite these features, our model can be reduced to a special case of CSP (Communicating Sequential Processes). Section 2 introduces new program design tools - shells - for the proposed programming languages. Section 3 considers some issues connected with objects and dynamic location. Section 4 describes the proposed method of program design. The paper concludes with remarks on implementation of the program execution system for the proposed programming languages.« less