Toward A Quantifiable Definition of Software Faults

• Authors: John C. Munson;Allen P. Nikora
• Venue: ISSRE '02 Proceedings of the 13th International Symposium on Software Reliability Engineering
• Year: 2002

An important aspect of developing models relatingthe number and type of faults in a software system to a setof structural measurement is defining what constitutes afault. By definition, a fault is a structural imperfection ina software system that may lead to the system's eventuallyfailing. A measurable and precise definition of whatfaults are makes it possible to accurately identify andcount them, which in turn allows the formulation of mod-elsrelating fault counts and types to other measurableattributes of a software system. Unfortunately, the mostwidely-used definitions are not measurable - there is noguarantee that two different individuals looking at thesame set of failure reports and the same set of fault definitionswill count the same number of underlying faults.The incomplete and ambiguous nature of current faultdefinitions adds a noise component to the inputs used in