Key points are not available for this paper at this time.
Experiments with quantitative assessment and prediction of software reliability are presented. The experiments are based on the analysis of the error and the complexity characteristics of a large set of programs. The first part of the study concerns the data collection process and the analysis of the error data and complexity measures. The relationships between the complexity profile and the error data of the procedures of the programs are then investigated with the help of discriminant statistical analysis technique. The results of these analyses show that an estimation can be derived from the analysis of its complexity profile.
Potier et al. (Mon,) studied this question.