Metrics for reusable software code components
Peter Fonash · 1993
This dissertation investigated a technical barrier to software code component reuse. The problem examined is the lack of a set of distinguishing attributes to automatically identify reusable components. To address the problem, static quantitative code metrics, based on software engineering principles (cohesion, coupling, parameterization, and quality) which correspond to reuse attributes and a review of the metrics literature, were developed. Using code components from the NASA Goddard Space Flight Center's Ada reuse library, measures for these metrics were collected from three reuse categories: reused without modification, reused after extensive modification, and new for application. Statistical analysis determined that ten of the measures differed significantly between the reused without modification and the reused after extensive modification categories. Fifteen measures differed significantly between the reused without modification and the new for application categories.