Regression analysis of reusable program components
Janusz W. Laski, Wojciech Szermer · 2002
An integrated regression analysis is proposed for the reverification of software components that are modified for a new application. The basis for the method is clustering, i.e., automatic identification of the modifications. The clustering is used to formulate a regression fault hypothesis about the existence of a fault in the new component and testing the hypothesis using static and dynamic (execution-based) analysis.>