QW2002 Paper 3A1

Dr. Erik Simmons
(Intel Corporation)

Software Defect Arrival Modeling With the Weibull Distribution

Presentation Abstract

One of the most common yet vexing questions asked of Software Quality Assurance managers and testers is "When will we be done testing?" Product engineering and marketing groups have a vested interest in knowing when the software under test will be at an acceptable level of quality. While this question is not at all easy to answer, modeling the arrival of defects during testing can provide clues such as predictions of when a given percentage of the estimated total defects will be found, or the time at which the rate of newly arriving defects will be below a given threshold. The Weibull distribution serves as an excellent model for software defect arrival. Three case studies based on actual projects are provided.

About the Author

Erik Simmons has 15 years experience in multiple aspects of software and quality engineering. Erik currently works as Platform Quality Engineer in the Platform Quality Methods group, part of the Corporate Quality Network at Intel Corporation. He is responsible for Requirements Engineering practices at Intel, and lends support to several other corporate software and product quality initiatives. Erik is a member of the Pacific Northwest Software Quality Conference Board of Directors and the Steering Committee of the Rose City SPIN. He holds a Masters degree in mathematical modeling and a Bachelors degree in applied mathematics from Humboldt State University in California, and was appointed to the Clinical Faculty of Oregon Health Sciences University in 1991.