You drive down the road and suddenly your car brakes to a stop—or accelerates uncontrollably. You try to withdraw money from your bank and find that your account is overdrawn, even though you think it should contain plenty of money. Your doctor phones to tell you a recent test showed that your usually normal vitamin D level is a fraction of what it should be. And your favorite candidate loses an election that should have been a sure victory. Should you be worried?
There may be other explanations for these events, but any of them may be the result of a computer security problem. Computers are embedded in products ranging from dogs to spaceships; computers control activities from opening doors to administering the proper dose of radiation therapy. Over the last several decades, computer usage has expanded tremendously, and our dependence on computers has increased similarly. So when something goes awry, it is reasonable to wonder if computers are the source of the problem.
But can we—and should we—depend on computers to perform these tasks? How much can we entrust to them, and how will we determine their dependability, safety, and security? These questions continue to occupy policy makers, even as engineers, scientists, and other inventors devise new ways to use computers.
From one perspective, these failures are welcome events because we learn a lot from them. Indeed, engineers are trained to deal with and learn from past failures. So engineers are well qualified to build large structures on which many of us depend. For example, consider bridges; these days, bridges seldom fail. An engineer can study stresses and strengths of materials, and design a bridge that will withstand a certain load for a certain number of years; to ensure that the bridge will last, the engineer can add a margin of safety by using thicker or stronger materials or adding more supports. You can jump up and down on a bridge, because the extra force when you land is well within the tolerance the engineer expected and planned for. When a bridge does fail, it is usually because some bridge component has been made of defective materials, design plans were not followed, or the bridge has been subjected to more strain than was anticipated (which is why some bridges have signs warning about their maximum load).
