Reviews: Thinking inside the Box; October 2003; Scientific American Magazine; by Michael M. Sokal; 2 Page(s)

Early-19th-century American steam engines were less fuel-efficient and more dangerous than their late-18th-century British counterparts. These details don't surprise most first-year engineering students. After all, didn't American technology lag Britain's for many years? They are often taken aback, however, to discover that these faults derived partly from explicit choices of American steam engineers. Historian George Basalla suggested in his 1988 book The Evolution of Technology (Cambridge University Press) that such choices can be understood as adaptations to the resource-rich and skills-poor American environment, in which heavy fuel consumption mattered less than the ease of design, construction and maintenance afforded by high-pressure operation. (These higher-pressure engines could also run at greater speeds, an important feature on a larger landmass.) More generally, like the Y2K problem, what later generations saw as a design flaw can be most richly seen as the result of a designer's attempt to work within the technological boundary conditions of a given time and place.

In Small Things Considered, Henry Petroski's approach to the question of "why there is no perfect design" is less evolutionary than Basalla's and reflects his own experience as a practicing engineer and a keen observer of the made world and of how people live in it. But like Basalla, Petroski continually emphasizes that all made things, both physical and social, are designed, whether consciously or unconsciously, and that anyone designing anything must work within a set of physical and social constraints. As he writes, in considering the design of chairs, "All designs must involve tradeoffs, if not in materials, then in function; if not in cost, then in fashion; if not in quality, then in proportion; if not in size, then in shape; if not in this, then in that." The design process is thus often labyrinthine, and successive compromises in response to specific constraints close off and open up different choice points later in the process. Indeed, even as designers "perfect" their creations, they usually both improve (in some ways) and impair (in others) what came before.