A lot of things just didn’t work this week. Microsoft had a colossal outage of its email and other online services. Air traffic control failed catastrophically, forcing a nationwide ground stop of aircraft. A riverboat sank in New Orleans after being hit by a loose tanker ship.
■ It’s tempting to look for sinister causes, though none have been identified. It will also be tempting to identify individual scapegoats, though the failures seem too large to belong to just one person.
■ The ultimate root cause, in each case, is almost certainly some kind of system failure. Not a system in the sense of a literal router design or ATC communications protocol or steering mechanism, but rather in the sense of “a way of doing things that leads to outcomes”. Some process involving people broke down along the way in each of these failures.
■ They’re unlikely to be photogenic problems, either: The safe money would bet on profoundly mundane and uninteresting causes. As the world grows more complex, our failures tend to become more specific -- caused by narrow specialties that only a few people in the world really understand. It’s a problem illustrated well by an XKCD comic titled “Dependency”, which depicts all modern infrastructure depending precariously on “A project some random person in Nebraska has been thanklessly maintaining since 2003”.
■ Finding out what really happened is enormously important in all three cases. Ultimately, the public deserves the knowledge -- and the solutions that should emerge from careful analysis, not merely a rush to put a headline answer on the board. “Human error” is easy to blame, but the ultimate cause is almost always a system.