DURHAM, N.C. — For an engineer, Henry Petroski seems strangely enthusiastic about failure.
Not his own, of course. Fear of failure is what sent him, with a bachelor's degree in mechanical engineering, to graduate school rather than to work, and then to a career of teaching and writing, not designing and building.
From his vantage point, failures in design and construction present perfect teaching opportunities. They are object lessons in the history and practice and beauty of engineering. "Failure is central to engineering," he said in an interview. "Every single calculation that an engineer makes is a failure calculation. Successful engineering is all about understanding how things break or fail."
So whether the subject is the building specs in "The Three Little Pigs," the development of the flip-top beverage can or the storage of nuclear waste (a current focus of his), Dr. Petroski thinks and writes in terms of failure. Failure looms even in "The Pencil," his 400-plus-page look at the invention, evolution, crafting and use of the writing implement whose points are so prone to breaking. The book was a surprise best seller.
Dr. Petroski, who is 64, has preached his gospel of failure in books, lectures and articles for publications as diverse as Forbes and American Scientist, where he has a regular column. In the process, he has amassed numerous honors and awards, including membership in the National Academy of Engineering. He has also achieved the status that a reviewer in the journal Science predicted for him after the publication in 1985 of his first book, a catalog of calamity called "To Engineer Is Human."
He is "the meistersinger of the guild."
That first book has on its cover a photo of a famous failure, the collapse of the Tacoma Narrows Bridge, a graceful span across Puget Sound. Its roadway was so narrow and light that it swayed and twisted even in 40-mile an hour winds. It collapsed in 1940, a few months after it opened, in a disaster famously captured on film.
According to Dr. Petroski, the lesson of that bridge is not that it failed, but that it was deemed invulnerable to failure, a judgment that is always a mistake.
Or take Frank Gehry's design for the Walt Disney Concert Hall building in Los Angeles, which Dr. Petroski describes in his latest book, "Success Through Failure," published this year by Princeton University Press. According to Dr. Petroski, the high gloss of one side of the building reflected so much light at a condo across the street that residents suffered blinding glare and 15-degree temperature increases until the offending wall was resurfaced in a matte finish. This problem is the kind of "latent failure" that emerges only when a design is in use.
And then there is the rolling suitcase Dr. Petroski's wife, Catherine, a writer, bought on a recent trip. She chose it because of its convenient design, he recalls. Only when she used it did she discover it does not roll smoothly when it is full. Moral: a device does not have to fail utterly to be a failure.
In designing and building, engineers calculate how components of their design must perform, and how much stress they can endure before they will give way, an analysis Dr. Petroski says they apply to tasks as varied as driving across a bridge and bending and unbending a paper clip. The paper clip exercise is one he often uses in Introduction to Structural Engineering, one of the classes he teaches at Duke, where he has appointments in both engineering and history.
The analysis of engineering's failures offers some good lessons, Dr. Petroski writes. For example:
¶Success masks failure. The more a thing operates successfully, the more confidence we have in it. So we dismiss little failures — like the repeated loss of a space shuttle's insulating tiles launchings — as trivial annoyances rather than preludes to catastrophe.
¶Systems that require error-free performance are doomed to failure.
¶Computer simulations and other methods of predicting whether components will fail are themselves vulnerable to failure.
¶Devices can be made foolproof, but not damn-fool-proof. This engineering maxim is one of Dr. Petroski's favorites.
¶Today's successful design is tomorrow's failure, in that expectations for technology are continually on the rise.
¶A device designed for one purpose may fail when put to another use. (Is it fair to call that a failure? Dr. Petroski smiled. "Good question," he said.)