A simple model for software reliability growth, originally suggested by Jelinski & Moranda, has been widely used but suffers from difficulties associated with parameter estimation. We show that a major reason for obtaining nonsensical results from the model is its application to data sets which exhibit decreasing reliability. We present a simple, necessary and sufficient condition for the maximum likelihood estimates to be finite and suggest that this condition be tested prior to using the model.
No takes yet. Share an insight, caveat, or question.
Littlewood et al. (1981) studied this question.
Synapse has enriched 2 closely related papers on similar clinical questions. Consider them for comparative context: