Pretty Good Technologies and Visible Disasters

On the movie screen and in television commercials, technologies are usually portrayed as visions of perfection, their parts seamlessly functioning without a hitch. In real life, however, most technologies are only pretty good or, if excellent, achieve that performance only at great price. Two Florida-based technologies, the space shuttle, symbol of America's technological prowess, and the state's voting system, now an icon of technological dysfunction, remind us of the limits of technologies, especially as they are influenced by the American political and economic system. Like the 1986 Challenger explosion in its time, the voting machine fiasco has become an object of intense speculation. How did things go wrong?

It might help to consider the differences between space shuttles and voting systems--the first an excellent technology, the second a pretty good one. A pretty good technology works adequately most of the time without major problems or investments, whereas an excellent technology requires significant resources to function properly. (Each shuttle launch demands one to three million person-hours and between four hundred million and eleven hundred million dollars, depending on how you count.) When either technology fails, the results can sometimes be spectacular, truly visible disasters.

Most technologies are not trouble free, but many problems are either minor or resolved by learning how to work around them. If you spend time with automobile mechanics, you learn that every car has its own peculiarities. The mechanics learn to work around the shortcomings. If you fly, you know that flight delays are common, so you bring an extra book. Even excellent technologies often require their users to work around problems. Well before the Challenger exploded on 28 January 1986, NASA engineers and managers knew the O-rings in the solid rocket boosters were seriously flawed. But they also knew the O-rings had functioned on the previous twenty-four flights within defined parameters, so NASA was confident it had the problem under control.

Why do we ever settle for pretty good technology? The answer is simple: The best is the enemy of the good. Achieving a greater guarantee of performance usually requires far more resources--money, training, people--than buyers or users are willing to provide. Thus, if you are a county commissioner facing conflicting demands for schools, road construction, and election preparations, which will you choose to meet, especially when taxpayers are reluctant to invest more? Even more important, pretty good systems suffice for most situations. Voting systems are an excellent example of a pretty good technology. Most of the time they work well enough because most elections are decided by margins greater than 2 or 3 percent. The inherent flaws and inaccuracies are less than the margin of victory. A more accurate system would not change the result, only raise the cost.

Since election day the regular media and the internet community (there is a wonderful dissertation to be done on the latter's role) have put Florida under a public microscope for two reasons. First, the extreme closeness of the vote has kept the presidential election uncertain for weeks. Second, the Palm Beach County butterfly ballot effectively cost both candidates an undisputed victory. This was a visible disaster that has fully exposed all the problems of voting.

To be sure, very close elections often occur, but mainly at the local level, so they rarely attract much attention. If the presidential election were not so close, these problems would not have been noticed. But now the public is discovering just how trouble-prone voting can be. The media have found numerous cases of miscast ballots, irregular procedures, and poorly performing machines and people. Irregularities, in short, are regular.

All technologies are part of society and reflect its values, although not always in obvious ways. The three main types of voting systems in the United States are an excellent example. Lever machines, the behemoths of voting technology, have not been manufactured for decades, but are still used by one-fifth of voters...