Waltzing with Bears

DeMarco and Lister on why software organisations avoid naming their risks and what that avoidance costs them. The book sets out a working risk management process covering identification, exposure, containment and the use of risk reserves, and it makes the case for expressing a delivery date as a distribution rather than as a single confident number. There is a chapter on the risks that turn up on nearly every software project, including specification breakdown, scope inflation and turnover. The authors are blunt that a project with no risks worth managing is a project not worth doing.