Confidence-based reliability and statistical coverage estimation

In confidence based reliability measurement we determine that we are nt least C confident that the probability of n program failing is less than or equal to c1 bound 8. The basic results of this approach are reviewed and several additional results introduced, including r/w adaptive sampling rheorem which shows how confidence can be computed when faults are correcred as they appear in the resting process. Another result shows how fo carry auf testing in parallel. Some of the problems of staristical testing are discussed and an nlternn~ive merhod for establishing reliability called statistical coverage is introduced. At the cmf of making reliability estimres that are relnrive fo a fault model, smtisrical coverage eliminares rhe need for output validation during reliability estimation and allows the incorporation of non-statistical resting results into the statistical reliability esrimtion process. Sratistical taring and statistical coverage are compared, and their relarionrhip wirh ondidoml reliability growth modeling approaches is briefly