This note summarizes conclusions from a three year study about why released software fails. Our method was to obtain mature-beta or retail versions of real software applications and stress test them until they fail. From an analysis of the casual faults, we have synthesized four reasons why software fails. This note presents these four classes of failures and discusses the challenges they present to developers and testers. The implications for software testers are emphasized.
|Why Software Fails (And How Testers Can Exploit It)||1.12 MB|