Within NASA, there is an increasing awareness that software is of growing importance to the success of missions. Much data has been collected, and many theories have been advanced on how to reduce or eliminate errors in code. However, learning requires experience. This article documents a new NASA initiative to build a centralized repository of software defect data; in particular, it documents one specific case study on software metrics. Software metrics are used as a basis for prediction of errors in code modules, but there are many different metrics available. McCabe is one of the more popular tools used to produce metrics, but, as will be shown in this paper, other metrics can be more significant.