Error density and size in Ada software

The author and her colleagues have completed the development of a large project written in Ada. The project team kept error reports, and the many modules ranged greatly in size. They analyzed module size to see if there was a relationship with module quality. They used error density-defects per thousand lines-as an inverse measure of quality: the lower the error density, the higher the quality. The author presents and discusses the results, which lend support to the hypothesis that there is an optimal, intermediate module size.<>