An Integrated Classification Scheme for Efficient Retrieval of Components

Abstract

Reuse is the key paradigm for increasing productivity and quality in software development. To be able to reuse software components, whether it is code or designs, it is necessary to locate the component that can be reused. Locating components, or even realizing they exist, can be quite difficult in a large collection of components. These components need to be suitably classified and stored in a repository to enable efficient retrieval. Four schemes had been previously employed, free text, enumerated, attribute value and faceted classification. Experiences revealed that individual classification schemes were unable to solve the problems associated with component classification. We required a combination of classification techniques to meet the problems with individual schemes and to improve retrieval efficiency. This research looked at each of the classification techniques above, and proposes a new method for classifying component details within a repository.