Predicting Field Defects Based on Software Test Results

From a customer's perspective, field defects are a measure of the software reliability and quality and serve as a key input for maintenance planning. In this paper we present an approach based on a predictor to predict field defects when the system does not exhibit reliability growth at the tine of release. The standard Software Reliability Growth (SRG) models cannot be applied in this case. Our approach is illustrated with the analysis of two defect data sets: the OpenBSD defect data and defect data from a telecommunications system.