Structural failures yield more usable engineering knowledge than successes, because a success only confirms that a design exceeded its loads by an unknown margin, while a failure reveals the precise limit and the specific error. Engineers therefore advance the state of the art by studying collapses, not by cataloguing structures that merely stayed up.
Every new structure is an implicit prediction that it will withstand forces and conditions not yet experienced, so designing is an act of forecasting under uncertainty rather than applying settled fact. The design is confirmed or refuted only when the real world tests it, which means some degree of guessing is inherent to the practice.
In 1940 the Tacoma Narrows suspension bridge in Washington State twisted and tore itself apart in a moderate wind only months after opening, because its slender, solid-sided deck interacted with aerodynamic forces engineers had not anticipated. The failure forced the profession to incorporate wind-tunnel testing and aerodynamic analysis into bridge design.
Repeated success tempts engineers to extrapolate proven designs to larger scales or lighter margins, eroding the safety buffers that earlier caution provided until a failure occurs. Catastrophes thus tend to follow periods of confident scaling-up, roughly every few decades, rather than striking randomly.
Ordinary engineered things, from paper clips to zippers to bridges, can be understood as accumulated responses to past failures, so their shapes and features encode a history of what went wrong before. Seeing an object this way reframes its details as solutions to problems rather than arbitrary choices.