A Framework for Effective Object-Oriented Software Change Impact Analysis

Object-oriented (OO) software have complex\uddependencies and different change types which frequently affect\udtheir maintenance in terms of ripple-effects identification or\udmay likely introduce some faults which are hard to detect. As\udchange is both important and risky, change impact analysis\ud(CIA) is a technique used to preserve the quality of the software\udsystem. Several CIA techniques exist but they provide little or\udno clear information on OO software system representation for\udeffective change impact prediction. Additionally, OO classes\udare not faults or failures-free and their fault-proneness is not\udconsidered during CIA. There is no known CIA approach that\udincorporates both change impact and fault prediction.\udConsequently, making changes to software components while\udneglecting their dependencies and fault-proneness may have\udsome unexpected effects on their quality or may increase their\udfailure risks. Therefore, this paper proposes a novel framework\udfor OO software CIA that allows for impact and fault\udpredictions. Moreover, an intermediate OO program\udrepresentation that explicitly represents the software and allows\udits structural complexity to be quantified using complex\udnetworks is proposed. The objective is to enhance static CIA\udand facilitate program comprehension. To assess its\udeffectiveness, a controlled experiment was conducted using\udstudents’ project with respect to maintenance duration and\udcorrectness. The results obtained were promising, indicating its\udimportance for impact analysis