Design of an interpretive environment for Turing

Abstract

This paper presents the design of an interpreter structure for modern programming languages such as Turing and Modula II that is modular and highly orthogonal while providing maximal flexibility and efficiency in implementation. At the outermost level, the structure consists of a front end, responsible for interaction with the user, and a back end, responsible for execution. The two are linked by a single database consisting of the tokenized statements of the user program. Interfaces between the major modules of each part are defined in such a way as to maximize reusability, and each interface can service a range of plug-compatible modules implementing radically different semantics. The design accommodates a wide spectrum of interpreter types ranging from batch-oriented compiler-simulators to statement-by-statement interactive execution, and provides for a range of program editing tools from simple line editors through to modern language-directed programming environments. It has served as the basis for several interpretive systems including the production Turing interpreter, the Turing Programming Environment, and the Turing Tool software maintenance tool.