I'm not prepared to write a category killer to supersede the whole node, but the writeup above glosses over critical information. So this will have to serve as a plug-in to the above writeup, until somebody writes the killer.
Murphy's Law is not a joke. Anything that can go wrong will go wrong eventually. Engineering is a profession of anticipating things that can go wrong, and designing systems to reduce or eliminate the consequences.
Although metal fatigue was the proximate cause for the sudden failures of two De Havilland Comets in 1954, it is a grave mistake to assume it was the fundamental cause. The fundamental failure was not in the materials; it was in the design, and the testing of that design.
If you look at photographs of early Comets, you will see something very strange: square windows. You never see these in modern planes, because it was the squareness of the windows that led to the crashes.
Investigations into the 1954 Comet crashes gave people in the 1950's a foretaste of modern airplane crash investigation, things we are (sadly) familiar with today: Cameras and divers sent to the bottom of the sea to investigate wreckage, and bits pulled up from the bottom for analysis.
Crash investigators subjected mockups of fuselage sections to simulated flight stresses, including flexing during takeoff, and differential pressures (the cabins were pressurized, so in the tests, they filled the cabins with water).
They discovered that stress loads in the aircraft's skin were concentrated around the window corners. Consequently, those corners were where the metal fatigue first turned into failure.
An Italian trawler pulled up the "smoking gun" from the January 10 crash off Elba: A section of cabin whose torn edge extended from a corner in a square navigation window.
This was the real blow to de Havilland's reputation: They put snazzy new features into production without fully testing their long-term effects. Even allowing for limited knowledge of metal fatigue at the time, the problems of windows with corners should have been anticipated. In fact, they were anticipated: The US Civil Aeronautics Administration never gave the early Comets airworthiness certificates because of their misgivings about the Comets' windows.
The Comet crashes inspired Australian engineer Dr. David Warren to invent the now familiar cockpit voice recorders and flight data recorders for civil aircraft.
To correct the metal fatigue problem, De Havilland put oval windows into later versions of the Comet, beginning a long tradition of tombstone technology for the aircraft industry.
Too late for de Havilland. Boeing, which lost the race to produce the first commercial jetliner, was able to learn from de Havilland's mistakes and produce the most successful commercial jet airliner, the 707. US airplane manufacturers dominated the industry for decades, only recently running into serious competition from the European consortium Airbus.
Century of Flight: deHavilland Comet
Engineering Disasters, The History Channel