Measuring the field quality of wide-distribution commercial software

The problem of quantifying the field quality of wide-distribution commercial software is addressed. The authors argue that for this type of software the proper quality metric is an estimate of the number of defects remaining in the code. They observe that the apparent number of defects remaining in this type of software is a function of the number of users as well as the number of actual defects in the code. New releases of commercial software normally consist of some code from prior releases and some new or modified code. The authors continue to discover new defects in the code from prior releases even after a new release is in the field. In fact, the new release code appears to stimulate discovery of defects latent in the 'old' code and cause a 'next release effect.' Field defect data from several releases of a widely distributed commercial software product are shown to a demonstrate this effect.<<ETX>>