Improving Fault Detection in Modified Code --- A Study from the Telecommunication Industry

Many software systems are developed in a number ofconsecutive releases. In each release not only new code is added butalso existing code is often modified. In this study we show that themodified code can be an important source of faults. Faults are widelyrecognized as one of the major cost drivers in software projects.Therefore, we look for methods that improve the fault detection in themodified code. We propose and evaluate a number of prediction modelsthat increase the efficiency of fault detection. To build and evaluateour models we use data collected from two large telecommunicationsystems produced by Ericsson. We evaluate the performance of our modelsby applying them both to a different release of the system than the onethey are built on and to a different system. The performance of ourmodels is compared to the performance of the theoretical best model, asimple model based on size, as well as to analyzing the code in arandom order (not using any model). We find that the use of our modelsprovides a significant improvement over not using any model at all andover using a simple model based on the class size. The gain offered byour models corresponds to 38~57% of the theoretical maximum gain.