| Software-based digital systems are progressively replacing analog systems in safety-critical applications like nuclear power plants. However, the ability to predict their failure rate is still not well understood and requires further study, despite the fact that there are over 100 software reliability assessment/prediction models in existence to date.; Software reliability is essentially determined by two factors: the product characteristics and the operational environment. The product characteristics are further determined by the project characteristics and the development characteristics. All above characteristics can be measured explicitly or implicitly using software engineering measures. Therefore an obvious inference is that “software engineering measures determine software reliability.”; A set of software engineering measures from which the software reliability can be predicted are identified. The concepts of structural classification, software development life-cycle classification, and family are presented. These measures are categorized using these concepts. The concept of RPS (reliability prediction system) and an extended structural representation are introduced to bridge the gap between software engineering measures and reliability. Expert opinion is elicited as the input in ranking the identified measures in terms of software reliability prediction. The top-ranked measures and families are presented. Six RPSs are constructed from 6 measures to assess the reliability of a small-scale system. The results validate the rankings and the RPS theory.; This study is the first step towards a systematic approach predicting the reliability of a real-time I&C software using RPSs established from the measures and families. |
