Fault prediction for software
John H R May, Patrick A. V. Hall, Haodong Zhu, Trevor Cockram, Nathan Bird, L.A. Winsborrow · Bristol Research (University of Bristol) · 1995
Most activities during the software development process have the potential to cause faults in the final software code. A fault caused somewhere in the development process will be manifested in products produced after its introduction e.g. a high level design document or a code component. A probabilistic model of fault introduction is presented, with the aim of estimating the integrity of the work carried out up to a particular point in a given development process. In particular, estimates of the integrity of products early in the development lifecycle are sought.