A defect tracking tool for open source software

Defect reporting and correction is one of the most crucial part in any phase of software development. This is a very costly activity. A lot of time, effort and resources can be saved if the defects can be predicted beforehand, using a suitable training data set. For this, the generation of defect reports which reports the classes that are defective and computes the required software metrics is required. This process should be fully automated. Various data analysis research techniques and algorithms can be applied on this produced data set and this information can be very helpful for the developers and the organization as a whole for the prediction of bugs for future purposes and for analysis of the quality of the software being produced. The various issues addressed in this paper are (1) selection of suitable projects as input for the tool, (2) a mechanism to trace the whole project code history and identifying the classes present, (3) the criteria for classifying a class as defective, (4) method to generate most accurate software metrics for the classes, (5) combining all the data into a final representable format. The tool is developed using Java, using the top open source repositories hosted on Github to test its working, incorporating SourceMeter tool to produce the software metrics and wrapping all this in a simple interface.