Abstract:
The paper presents and discusses an industrial case study, where an eight year running software project has been analyzed. We collected about 1000 daily-versions, together with the file version control system, and bug tracking data. This project has been migrated from Java 1.4 to Java 1.5, and visible effects of this migration on the bytecode are presented and discussed. From this case study, we expect to observe the effects on the code size produced by the Java technology migration, and to improve the performances of already existing fault-proneness estimation models. Preliminary results about fault-proneness estimation are shown.
