Value Driven Testing

The following contribution is an extension of the latest research on value-driven software engineering to the field of software testing. The goals and methods of value driven software engineering are reviewed and analyzed in regard to their application to testing. The particular difficulties of cost justifying testing are pointed out. The point is made that despite these difficulties it is possible to quantify the benefits of testing in terms of the costs of errors occurring in production and the costs of correcting those errors. The paper proposes using the methods of risk analysis to predict those costs. To predict test costs a modified COCOMO-II equation is proposed. In the end the costs and benefits are compared to compute a return of investment - ROI - on testing.