A FRAMEWORK FOR SOFTWARE DEVELOPMENT PRODUCT METHODOLOGY MANAGEMENT

ABSTRACT: A software systems developer or manager needs to have an overall visibility of the infrastructure of a system and needs to have a clear picture of all the relationships between the various components of the system. This paper describes methods for enhancing visibility which can be provided automatically by extracting the required information from the coding and testing phases and from the related software development phases. The extracted items of information are recorded into a knowledge base structure. The research in this paper furnishes a software developer or manager with capabilities that enable one to establish quickly and efficiently the correspondence between software components and the required information from other phases of the software methodology. This information, therefore, can support and enhance decision making through the complete life cycle of a software product. The paper describes a method of analyzing software programs whereby cross-referential, dependency, and other information can be abstracted automatically during compilation to support developers and /or managers to adopt the appropriate decision-making. These items of information are also used to support library administration, configuration management, version control, software reusability, and software composition. The analysis process is accomplished by transcribing underlying language grammar rules directly into programming language Prolog as predicates in first-order logic. The research provides a data querying mechanism that enables developers and managers to investigate the logical basis for improving abstraction, maintainability, and usability of the software methodology. This is accomplished through the browsing of items of data stored in the underlying knowledge representation systems, especially the improvement of the querying answering facilities of such systems. A knowledge representation system allows the formulating and recording of constraints and facts about a domain of the real world in a knowledge base. A knowledge representation system provides inference services which enable deducing implicit facts. With queries, developers and/or managers can directly or indirectly deduce and retrieve knowledge stored in the underlying knowledge base.