Software reliability measurement

  • Authors:
  • John D. Musa

  • Affiliations:
  • Bell Laboratories, Whippany, New JerseyU.S.A.

  • Venue:
  • Journal of Systems and Software
  • Year:
  • 1984

Quantified Score

Hi-index 0.00

Visualization

Abstract

The quantification of software reliability is needed for the system engineering of products involving computer programs and the scheduling and monitoring of software development. It is also valuable for the comparative evaluation of the effectiveness of various design, coding, testing, and documentation techniques. This paper outlines a theory of software reliability based on execution or cpu time, and a concomitant model of the testing and debugging process that permits execution time to be related to calendar time. The estimation of parameters of the model is discussed. Application of the theory in scheduling and monitoring software projects is described, and data taken from several actual projects are presented.