An Early Look at COQUALMO in the JPL Environment

This paper will describe the initial efforts to validate and calibrate the COQALMO model within the JPL environment. The risk to JPL software projects precipitated by higher than expected defect densities late in the software lifecycle impact spacecraWmissions in several ways. Among them are risks of missing “hard” schedule deadlines, which may be dictated by immovable launch dates and the timing of critical spacecraft encounters with objects of interest in space. Further risks include, but are not limited to, failures during execution of mission critical software that has only one opportunity to perform its functionality correctly. For example, software critical to mission critical nonrepeatable phases such as Entry Decent Landing (EDL) software and instrument and data collection software during mission encounters. The ability to predict expected defects densities early in the life cycle with models such as COQUALMO will allow project managers to plan schedules and resources more efficiently to ensure the delivery of high quality software within the critical time constraints discussed above. However, as with any predictive statistical model like COQUALMO, it must be calibrated to the environment in which it is to be used (i.e. JPL) to provide maximum benefits to software projects that come to rely on the information it provides.