DOES THE iHGOLDILOCKS CONJECTURElS APPLY TOSOFTWARE REUSE? AN EXPLORATORY STUDYUSING A DOMAIN-SPECIFIC REUSE MODEL

The Goldilocks Conjecture represents a speculation that there exists an optimal module size and is based upon behavior observed in software defect prediction models. While some data sets suggest that the Goldilocks Conjecture may apply to defect density prediction, the effect of module size on software reuse effectiveness is not fully understood. Breakeven analysis represents an important aspect of software reuse. This paper examines the breakeven point for repositories of reusable software modules, in relation to both module size and repository size, using a domain-specific model of software reuse. The findings suggest that a variant of the Goldilocks Conjecture may apply, but prescriptions for software reuse practitioners must be made with caution.