Wednesday, May 04, 2005
John Thackara, in his absolutely brilliant book, In the Bubble, talks extensively about the unintended consequences that numerous micro-decisions can form in over time. Thackara is careful to distinguish his concern for the consequences of disfunctional behaviors, from a comdemnation of the actors involved. He rightfully notes, as very few commentors do, that it is not greed or stupidity that causes most disfunction, it is the embedded rules that shape human behaviors.
I'm only half way through the book, but I can say it is one of the most significant I have read in some time.
Unfortunately, it is easier to lament disfunctional design than to understand it. Thackara is very unique is being able to do this. Others only scream at problems. A recent editorial article in Metropolis, How Piecemeal Design Approaches Hurt Us, provides such an example. The author blasts the engineers of computers for making them run too hot. What idiots, she protests, don't they know how unpleasant hot computers are? I'm not an engineer, but I have known quite a few, and I'm fairly certain they are desparate to cool their computers. The saniety of the unhappy editorial writer is the least of their motivations. Hot computers are bad business: they cost more to run, and breakdown more quickly as well. Thackara notes how large multinationals are very sensitive to waste: it is a sign of inefficiency.
Before we dismiss piecemeal design as a cardinal sin, we need to consider its positive qualities as well. The entire open source movement is an example of piecemeal design. I feel it has yielded some amazing benefits, though I wouldn't be surprised if some downside of it is revealed in due time. Much innovation in history has been piecemeal, and there is often a mixed outcome involved. We have become more sophisticated at understanding the difficulty of controlling outcomes. But we haven't become more sophisticated at controlling them. Some very high level systems thinking is needed, and some common understandings and resolve.