Abstract

Based on a simple non-procedural language with temporal logic operators, Lucid underlies a family of multi-dimensional programming languages based on intensional logic. Intension is a concept rooted in an aspect of natural language called "intensional context", in which the meaning of a statement (extension) depends on the context in which it is uttered (intension). The implicit temporal feature of Lucid makes it suitable for use as a means of describing dynamic systems. In the past, experiments have been performed and real applications have been developed with programs written in Lucid. However, these systems focused mainly on improving the execution performance of one dialect of Lucid and not address the problem of interpreting variants of Lucid. The GIPSY system is designed to not only process current Lucid variants efficiently but also to be modified easily to accept new dialects of Lucid. In the thesis, we discuss the essence of executing intensional programming languages using the eduction (also called demand-driven or lazy ) execution model; describe experiments with different approaches to interpreting programs written in Lucid; and focuses on execution over a network of processors. We describe the implementation of a prototype for executing Lucid programs in a distributed environment. We also explore the advantages of applying the object concept to distributed systems and describe experiments with these methods. In addition, the thesis includes estimates of the impact of integrating computation functions into the Lucid code and proposes an advanced execution model consisting of self-contained and intelligent clients associated with a meta-level resource management.