An improved tool for automated compiler construction

Since the advent of High Level Programming languages (HLPLs) in the early 1950s researchers have sought ways to automate the construction of HLPL compilers. To this end a variety of Translator Writing Tools (TWTs) have been developed in the last three decades. However, only a very few of these tools have gained significant commercial acceptance. This thesis re-examines traditional compiler construction techniques, along with a number of previous TWTs, and proposes a new improved tool for automated compiler construction called the Aston Compiler Constructor (ACC). This new tool allows the specification of complete compilation systems using a high level compiler oriented specification notation called the Compiler Construction Language (CCL). This specification notation is based on a modern variant of Backus Naur Form (BNF) and an extended variant of Attribute Grammars (AGs). The implementation and processing of the CCL is discussed along with an extensive CCL example. The CCL is shown to have an extensive expressive power, to be convenient in use, and highly readable, and thus a superior alternative to earlier TWTs, and to traditional compiler construction techniques. The execution performance of CCL specifications is evaluated and shown to be acceptable. A number of related areas are also addressed, including tools for the rapid construction of individual compiler components, and tools for the construction of compilation systems for multiprocessor operating systems and hardware. This latter area is expected to become of particular interest in future years due to the anticipated increased use of multiprocessor architectures.

Economic aspects of safety in the Greek construction industry

The thesis addresses the economic impacts of construction safety in Greece. The research involved the development of a methodology for determining the overall costs of safety, namely the sum of the costs of accidents and the costs of safety management failures (with or without accident) including image cost. Hitherto, very little work has been published on the cost of accidents in practical case studies. Moreover, to the author’s belief, no research has been published that seeks to determine in real cases the costs of prevention. The methodology developed is new, transparent, and capable of being replicated and adapted to other employment sectors and to other countries. The methodology was applied to three construction projects in Greece to test the safety costing methodology and to offer some preliminary evidence on the business case for safety. The survey work took place between 1999 and 2001 and involved 27 months of costing work on site. The study focuses on the overall costs of safety that apply to the main (principal) contractor. The methodology is supported by 120 discrete cost categories, and systematic criteria for determining which costs are included (counted) in the overall cost of safety. A quality system (in compliance with ISO9000 series) was developed to support the work and ensure accuracy of data gathering. The results of the study offer some support for the business case for safety. Though they offer good support for the economics of safety as they demonstrate need for cost effectiveness. Subject to important caveats, those projects that appeared to manage safety more cost-effectively achieved the lowest overall safety cost. Nevertheless, results are significantly lower than of other published works for two main reasons; first costs due to damages with no potential to injury were not included and second only costs to main constructor were considered. Study’s results are discussed and compared with other publish works.