To Engineer Is Human: The Role of Failure in Successful Design - Henry Petroski Perhaps I rate this too highly. Problem is I love technology and its issues and Petroski is one of my favorite writers on civil engineering.

On the 50th anniversary of the Golden Gate Bridge, May 27, 1987, almost 1,000,000 people showed up to celebrate and to walk across a bridge that was designed using the same basic technology as the infamous Tacoma Narrows bridge. Only about 250,000 were able to squeeze on the bridge, and fortunately no panic occurred as the Golden Gate Bridge began to sway gently from side to side. Hangar cables became slack -- something that was not supposed to happen, and the main span's arch flattened out to a "noticeable degree." The bridge had been over-designed with an ample margin of safety, unlike the walkways at the Hyatt in Kansas City, which were essentially small bridges. Over 100 people were killed when the walkways collapsed. Engineers determined quickly that a change made to make installation of the walkways simpler reduced the ability of the walkways to handle even their own weight let alone that of several hundred people.
Henry Petroski, in To Engineer is Human: The Role of Failure in Successful Design, is interested in engineering failures. He suggests these are terribly important to study, for they provide the clues to resolving the inherent paradox in engineering, which is that "...successful structural concepts devolve into failures, while the colossal failures contributed to the evolution of innovative and inspiring structures."

Structures that never fail -- actually they all will eventually, if one takes them beyond their intended life -- are assumed to be over-designed, i.e., they are much stronger than need be. Engineers, in order to be more economical and aesthetic, will make changes in the design that may ultimately lead to sensational failures like that of the Tacoma Narrows bridge. It's designers ignored considerable evidence that was readily available on the effect of wind on non-stiffened structures.

Petroski is concerned that the current atmosphere of liability and law suits will lead to a suppression of free discussion of the reasons behind structural (and now computer program) failures. "Engineering is a human endeavor and thus subject to error." Catastrophes are rare, but Petroski discusses why failures may be impossible to avoid and also why, paradoxically, we may not want to make them impossible.