Selecting an appropriate model for software reliability

Three models which have distinct approaches to predicting software reliability are discussed. Model A, the mean time to failure (MTTF) error-correction model, implements the viewpoint of reliability based on the number of software errors corrected. Model B, the MTTF error-detection model, uses elapsed testing time and the number of errors detected as the basis for the prediction. Model C is a simulation of the seeding model. Each of these models predicts the number of residual errors in the system to calculate the MTTF of the software. These models were selected prior to the start of testing because the assumptions and data necessary for each model closely fit those of the existing testing environment