Validity of Execution-Time Theory of Software Reliability

This paper investigates the validity of the execution-time theory of software reliability. The theory is outlined, along with appropriate background, definitions, assumptions, and mathematical relationships. Both the execution time and calendar time components are described. The important assumptions are discussed. Actual data are used to test the validity of most of the assumptions. Model and actual behavior are compared. The development projects and operational computation center software from which the data have been obtained are characterized to give the reader some basis for judging the breadth of applicability of the concepts.