A stochastic model of fault introduction and removal duringsoftware development

摘要

Two broad categories of human error occur during softwarendevelopment: (1) development errors made during requirements analysis,ndesign, and coding activities; (2) debugging errors made during attemptsnto remove faults identified during software inspections and dynamicntesting. This paper describes a stochastic model that relates thensoftware failure intensity function to development and debugging errornoccurrence throughout all software life-cycle phases. Software failurenintensity is related to development and debugging errors because data onndevelopment and debugging errors are available early in the softwarenlife-cycle and can be used to create early predictions of softwarenreliability. Software reliability then becomes a variable which can bencontrolled up front, viz, as early as possible in the softwarendevelopment life-cycle. The model parameters were derived based on datanreported in the open literature. A procedure to account for the impactnof influencing factors (e.g., experience, schedule pressure) on thenparameters of this stochastic model is suggested. This procedure isnbased on the success likelihood methodology (SLIM). The stochastic modelnis then used to study the introduction and removal of faults and toncalculate the consequent failure intensity value of a small-softwarendeveloped using a waterfall software development