Bach vs. Microsoft

What does a Bach fugue have in common with the Panama Canal? And how do they both differ from an oil rig or a software program?

If you have ever attempted to learn a Bach fugue, you have almost certainly been overwhelmed at times by the sheer implausibility of it. It doesn't seem possible that the composer could have kept it all in his head to create it in the first place, let alone that anyone could perform it with the skill and grace that the music demands. Unless you are among the few who have been endowed with an exceptional musical talent, you have probably had to come to terms with never being able to play the piece the way it is meant to be played.

Meant to be mastered

Even the truly gifted must wonder at how Bach managed to adorn extraordinarily complex structures with such alluring and easily perceptible lyricism. Despite the challenges, the fugues have indeed been played brilliantly by quite a few enviable artists, along with many other Bach creations of surpassing beauty and maddening difficulty.

Not all of man's inventiveness is intended for the glory of art (or man or God). Some is purely utilitarian, and still defies simple comprehension. Consider an off-shore oil platform. A government commission released a report recently on last summer's oil spill in the Gulf of Mexico. The authors of the report made a point of reminding a deeply skeptical public that they were charged with studying the conditions that existed on the oil platform at the time of the accident, not on any other considerations that may have contributed to the worst oil spill in our history. What those other factors may have been went unmentioned, but the conflicting interests and accounts of the major parties — the oil company that leased and operated the rig, the company that built the rig, and the oil services company that provided the cement that was deemed to have played a critical role in the well's failure — plainly demonstrate that the enormity of designing, building, deploying and operating an oil platform makes deep-sea drilling staggeringly complicated, and therefore implicitly and unavoidably dangerous.

It is not simply a matter of having so many entities involved that no one of them can take responsibility for the overall operation. The problem is more fundamental, and is emblematic of our age. We have grown so accustomed to grandiose feats of engineering that we have come to believe that, given enough time, money and brainpower, anything humans set out to build can indeed be built successfully. Notwithstanding the evidence of our own eyes — explosions at oil wells, the collapse of bridges, the failure of the dikes in New Orleans — humans seems to have a natural tendency towards greater and greater achievements of construction. While the industrial age has added breadth and velocity to our exploits of engineering, the quest is hardly new. As soon as a human population assembles in sufficient quantity to provide a large army of laborers, the building soon follows. The ancient Egyptians, Mayans, Greeks, Romans and many other early civilizations have left their marks in stone across the landscape. Steel and concrete may be products of the modern age, but the mathematics necessary for large-scale projects was discovered thousands of years ago. We ought not to marvel, then, as we often do, at how the ancients were able to build their great monuments, as if they were any less intelligent or resourceful than we are. The perfect acoustics of an ancient amphitheater is sufficient evidence of the ingenuity of our ancestors. What we should not imagine, however, is that they were any better at building than we are. For every extant Gothic cathedral there is at least one other that ended in disaster. Even many of the surviving ones, remarkable and exquisite though they are, have experienced deadly partial collapses, and would not still be standing without the interventions of engineers in later centuries.

In the information age, our building instinct has turned largely towards the making of systems. Here, too, our faith in technology constantly runs up against the unavoidable reality of its frequent failure. We are often exasperated, but never surprised, when the computer system at a bank or an airline is "down", or when our home computer or mobile device starts behaving badly. Somehow, that does not seem to dampen our enthusiasm for the next great gadget. Without any meaningful statistical analysis to guide us, we are nonetheless convinced that the rate of success and the promise of advancement outweigh the risk of failure. We cannot seem to help ourselves.

There are indeed innumerable successes to maintain our faith in man's ceaseless ingenuity. For example, there has not been a single major calamity at the Panama Canal in nearly a century of continuous use (apart from the tens of thousands who lost their lives during its original construction). If you were to look at photographs of the building of the canal, or watch one of the documentaries about it, you would likely react in much the same way as a pianist responds to the challenge of a Bach fugue: with sheer astonishment that anyone could have even conceived of something so complicated.

Let's return, then, to the original question: what do the fugue and the canal have in common? Both emerged from the imagination of a single individual. Unlike the fugue, the creation of which did not require Bach to employ a legion of helpers, the canal demanded the expertise of many people. But by and large it was the brainchild of the American engineer John Frank Stevens, who rejected the sea-level idea of Ferdinand de Lesseps in favor of the now-familiar but immensely more complex system of dams and locks. Stevens obviously could not have done all the calculations and drafted all the drawings by himself, but it is not far-fetched to suggest that he did indeed conceive of the entire project in his mind. The canal works as a totality, at least in part, because it was conceived and built as one object, like the Great Pyramid or the Eiffel Tower, and not as a contraption assembled from the ideas of numerous people working separately.

Most large software programs have the unmistakable feel of something built by a committee. The most complicated applications have so many features that no one person can ever learn them all. An oil platform, similarly, looks like something cobbled together from thousands of disparate parts, which in fact it is. These modern systems cannot be mastered entirely by any individual because they were not created by an individual. No one manager can reasonably be expected to get his head around everything that can possibly go wrong at a nuclear power plant or in a sprawling computer network because the task is, by definition, humanly impossible. This is not to suggest that the conceptions of single individuals cannot also fail, because they sometimes do. Or to suggest that systems cannot work, because they obviously do to a great extent. Jet aircraft take off and land uneventfully thousands of times a day. But even shining successes like manned flight and the space program have had more than their share of accidents and fatalities. Any risk of failure greater than zero is higher than we would prefer to entertain. Moreover, the consequences of those failures, like the gush of oil or the breaching of the dikes, increase in proportion to the immensity of the challenges these gigantic systems are intended to overcome.

Perhaps man ought not to build anything that is beyond the comprehension of at least one person, even if that person must be a virtuoso to master it. No one dies when a pianist misses a note, which is more than we can say about the next inevitable calamity.