Use case based approach to analyze software change impact and its regression test effort estimation

Abstract

Software needs to be maintained and changed to satisfy the new requirement and existing faults. Without analyzed changes, change implemented to software often cause unexpected ripple effects. To avoid this and diminish the risk of performing undesirable changes, an impact analysis of the change is done. Software Change Impact Analysis (SCIA) needs to be computed at every change request for software systems, to access the impact information for several critical software engineering tasks such as risk analysis, effort estimation, and regression testing. The use of UML analysis/design models on large projects lead to a large number of interdependent UML diagrams. Paper proposes a UML model based approach strictly to use use-case diagram for impact analysis that is applicable in early decision making and change planning. Later, by using the impact set we estimate the regression test effort required for the effected change in the software. The reduction in test effort observed ranges from 47 to 95 %, saving significant software testing cost.