(A) management discipline of software metrics and its application to database management systems

Abstract

There are still many problems to place software metrics on the center of mainstream software engineering. Currently, software engineers face the difficulty of quantification that helps software engineers in explaining cleanly what they have done and what attributes of their products the software metrics characterize. There are too much subjectivity and fuzziness both in software measurement methodology and in measurement results. Within the domain of software engineering, software metrics are understood as an unstructured items to do with measuring and predicting software project costs, productivity, quality and complexity of software products. That is, no rigorous engineering discipline is established for software metrics. In order to provide engineering disciplines of software metrics, this dissertation aims at solving their two problems: the management and utilization of knowledge required to provide consistent measurement environments. Specific problems are controlled management of software measurement activities and incomplete and inconsistent measurement environment. The management problem has produced a new discipline of software measurement, a management discipline of software metrics. Then, the software quality manager is proposed as a realization of the management discipline of software metrics. The proposed software quality manager is a generalized system that must prove its feasibility and applicability by applying to real software systems. In the software quality manager, it is essential to utilize the domain-specific knowledge about software systems to their measurement. Because of their high reusability, database management systems are chosen as an application area of the software quality manager. In software reuse techniques, software components are used to represent software in a modular way. Thus, information processing components are proposed in order to store their functional knowledge. The intra-structure of information processing component...