Abstract:
Generalizing components for future reuse is advocated
as a worthwhile investment of time and resources. Some
studies have attempted to quantify the additional costs of
generalization of object-oriented components for reuse,
with varying success. In this pilot study, we not only
analyze one small data set for values of generalization
egort but, more importantly, we develop an algorithmic
approach that can be used to predict how much data is
needed before a reliable estimate can be made.
