Discusses the need for quantitative descriptions of software errors and methods for gathering such data. The software development cycle is reviewed, and the frequency of the errors that are detected during software development and independent validation are compared. Data obtained from validation efforts are presented, indicating the number of errors in ten categories and three severity levels; the inferences that can be drawn from these data are discussed. Data describing the effectiveness of validation tools and techniques as a function of time are presented and discussed. The software validation cost is contrasted with the software development cost. The applications of better quantitative software error data are summarized.