With Complexity, Decks Are Stacked
Originally in Newsday, February 3, 2003
Almost before the mourning, the search for the explanation begins. When a public disaster like Saturday's space shuttle crash takes place, it's our natural impulse to find out why - an impulse motivated largely by a desire to avoid such tragedies in the future and to learn from our mistakes. Was it tiles damaged at takeoff? The wrong angle at rollover? A fuel leak? Insufficient funding?
Sometimes, this search can slide from inquiry to inquisition, and from putting our finger on answers to pointing it in accusation or blame. We surely don't know what happened now; perhaps, through the painstaking examination of debris and data, we will be able to trace the chain of events that happened Saturday morning, 200,000 feet above the southwest United States. We like our answers to these questions to be simple and unified. But in cases like this one, in the complex and unknown region at the boundaries of what we can do, they might well turn out to be neither.
In general, as systems get more complex, they tend also to become more susceptible to tiny failures. A pile of cards can have one pulled out without significant consequence, but a house of cards, where each does a particular job, is far less resilient. And in a highly complex system, those small failures can easily snowball into large-scale failures. Take computers, where a single "1" instead of "0" can mean the difference between crashing and working.
The shuttle is, of course, the epitome of complexity: 2.5 million parts, including the 27,000 tiles that protect against the heat of re-entry, combined with the almost unimaginable extreme conditions of speeds more than 17,000 miles per hour and temperatures of 2,800 degrees Fahrenheit, give us exactly the kind of situation where what works on paper has the least guarantee of working in practice. The tiniest variations may have the most profound effects. And without significant real-world experience, we won't know which of those tiny variations will occur, or which will have profound effects - a tile here or there, a degree or two of approach alignment, or, as with the Challenger in 1986, a slight drop in the temperature of the O-rings.
Surely there's a cautionary tale here about taking for granted our complex technologies - the shuttle, or more mundanely, the Internet or nuclear power. But before we rush to a Luddite's conclusion, it's important to see that this isn't just about technology. It's a general fact about complexity, and not just a particular problem for complicated mechanisms that we might build. Nearly every moment of our everyday lives rests on such delicate not-quite-catastrophes.
Consider natural systems, like the human body. It's part of the nature of complex interdependent biology that different jobs are done by different tiny parts of the system, and that those parts can be vulnerable to small problems, which can then amplify into large ones. The delicate subtlety required for the normal functioning of cells is also what makes them vulnerable to viruses.
In fact, it's in our natural world where the delicate balances are often the most difficult to grasp - for example, understanding the fragile balance of normal cellular life and growth that straddles the catastrophe of cellular death on one hand, and uncontrolled cancerous cellular multiplication on the other. In the end, we can't really avoid these delicate and sometimes catastrophic balancing acts, although we can sometimes manage and minimize them. And technology is sometimes the route to minimizing those catastrophes in both occurrence and consequence, as in much of modern medicine.
Our lives, and not just those of the shuttle crew, are often balanced on such chaos, as in the line between normal cell growth and cancer. We can't avoid all this by pulling back from technology. It assaults us all the time, even in our mundane lives. We're perched on precarious bits of marginally stable chaos that are surrounded by catastrophe, like a house of cards - cells and cancer, the working of the brakes on the truck behind us, and so on. The fragility of our lives and world might be managed, but never eliminated.
When we extend our reach, as with the space program, the uncertainties and complexities multiply, and the fragility of it all comes to the surface again. And whatever concrete explanation we may find in the months to come, we should not make the mistake of thinking we have made this activity completely safe. Let us not forget again, as we did before the Challenger in 1986 and perhaps had begun to do again, that the people who go up are test pilots, shouldering the risks for the rest of us in the worthwhile quest to expand the frontiers of knowledge.
Copyright © 2003, Newsday, Inc.