We introduce a general uniform language-independent framework for designing online and offline source-to-source program transformations by abstract interpretation of program semantics. Iterative source-to-source program transformations are designed constructively by composition of source-to-semantics, semantics-totransformed semantics and semantics-to-source abstractions applied to fixpoint trace semantics. The correctness of the transformations is expressed through observational and performance abstractions. The framework is illustrated on three examples: constant propagation, program specialization by online and offline partial evaluation and static program monitoring.

. The aim of this work is to compute descriptions of successful computation paths in logic or functional program executions. Computation paths are represented as terms, built from special constructor symbols, each constructor symbol corresponding to a specific clause or equation in a program. Such terms, called trace-terms, are abstractions of computation trees, which capture information about the control flow of the program. A method of approximating trace-terms is described, based on well-established methods for computing regular approximations of terms. The special function symbols are first introduced into programs as extra arguments in predicates or functions. Then a regular approximation of the program is computed, describing the terms occurring in some set of program executions. The approximation of the extra arguments (the trace-terms) can then be examined to see what computation paths were followed during the computation. This information can then be used to control both off-l...

Program specializers improve the speed of programs by performing some of the programs' reductions at specialization time rather than at runtime. This specialization process can be time-consuming; one common technique for improving the speed of the specialization of a particular program is to specialize the specializer itself on that program, creating a custom specializer, or program generator, for that particular program.

Abstract not found

The main contribution of the thesis is to develop a comprehensive framework for developing logic programs using #ve program schemata and some basic data types, and to evaluate its suitability for developing sizable programs. The schemata are intended to capture design decisions made by a programmer during top-down re#nement of a predicate. The schemata are completely independent from the data types, and have associated types and modes. The method of construction is to start from an initial typed moded call, and proceed using precisely defined refinement operations until primitive data type operations are introduced. It is shown that the construction method produces programs that are polymorphic many-sorted formulas, and that computations using the programs cannot result in type or mode errors. The framework is compared with previous schema-based program development methods from the literature, and we argue that this method is better suited to the developmentof non-trivial programs, and is more amenable to computerised support. The introduction of types and modes into schemata is a significant addition, not previously reported. An approach to program veri#cation is also introduced, in which the proof of correctness of a program is built from proof schemata corresponding to the program schemata used in the design of the program. Finally an implementation of the system is described and its use is illustrated on some sizable programs. The implementation incorporates a static analysis algorithm for checking the validity of the modes during the development of the program.