Bill Porteous argues that it is by discovering the causes of failure that progress may be made in reducing or controlling the incidence of building failures in the future — the paper proposes a way to describe the causes so that any failure event can be classified by cause.