Revisiting the debate: Are code metrics useful for measuring maintenance effort?

Evaluating and predicting software maintenance effort using source code metrics is one of the holy grails engineering. Unfortunately, previous research has provided contradictory evidence in this regard. The debate still open: as a community we are not certain about relationship between impact. In study investigate whether can indeed establish at previously unexplored method level granularity. We consider $\sim $ 730K Java methods originating from 47 popular open projects. After considering seven change proneness indicator, demonstrate why past studies contradict another while examining same data. also show that evaluation context king. Therefore, future should step away trying to devise generic models develop account for indicator being used size analyzed. Ultimately, be applied reliably these provide insight into when they judiciously context-sensitive manner.