Aiaa-99-4456 Practical Issues in Estimating Fault Content and Location in Software Systems

Over the past several years, techniques have been developed to discriminate between fault-prone software modules and those that are not, and to estimate a software system’s residual fault content. These techniques can be applied during the stages of a development effort prior to test, thereby giving software managers greater visibility into the systems being developed and allowing them to exert more accurate and precise control into their quality. There are practical issues involved in implementing these measurement techniques in a production development environment. In this paper, we describe measurement techniques being implemented on a development effort at the Jet Propulsion Laboratory, identify implementation issues, and describe proposed resolutions to these issues.