Could also have been called 'The Tao of Building' or the book that launched a fleet of pattern books.
I bought it hoping to get a better understanding of DesignPatterns, both the book and the patterns movement in software. It's strange that a book about architecture should be so influential in software. However, having read the book, I think the ideas apply more strongly to software than they do to buildings.
For me, the first few chapters were exhilerating. His explanation of the quality without a name reminds me of a book I always loved, ZenAndTheArtOfMotorcycleMaintenance, which also grappled with the question of quality near the start.
Others have mentioned this about the second book in the series, APatternLanguage, but the photographs in this book too are more than illustrations. They are striking and they relate to the topic of the book in a deep way. Checking at the back of the book, I found that no fewer than four are credited to HenriCartierBresson?.
I became less happy when it came to the second concrete example of the use of a pattern language: a clinic complex designed with the help of its users. ChristopherAlexander? is apologetic about the photograph, saying that the project did not turn out to be so great after all. He believed the initial design lost its qualities when it was turned into classic architectural drawings and built with standard materials.
I could not help thinking about the expense. He was able to take the entire staff for the best part of a week and have them walk around the site placing markers. No doubt about it, there's probably no better way to make sure their needs are met and that they 'buy in' to the concept. But if the salary clock is ticking that's a lot of money.
Going by the book, standardized materials where all the windows, doors, etc. are built to standard sizes cannot be alive, because they do not take into account the specific forces in their contexts. I may have a particular problem with this living in Chicago, but I wonder: would all the buildings here have to be discarded? For that matter, what do pattern languages have to say about buildings that are more than four stories tall? There is a photo of a model of a multi-family four story building looks utterly unconvincing, at least to me.
The idealism does become frustrating after a while. The book was written in 1979 and clearly, in twenty five years, most of it has not come to pass. We still commute to work, even though that makes whole neighbourhoods less alive. Power has not devolved down to communities and the city planning still fails to grow from the grassroots up.
It's a book of its time. Like ThePsychologyOfComputerProgramming, it commands us to become egoless. I don't know if it comes from Zen Buddhism, Taoism or just vague Eastern mysticism, but there is something of that flavour towards the end. It turns out that a pattern language is only a means of coping until we have learned to go beyond the polluting images we have in our heads. In the end, we can violate patterns as long as we balance the forces that are actually present.
I still have mixed feelings about the way the book appeals to scientific reasoning but then uses tools that are far removed from the normal tools of science. I may not know what the force of gravity is but I can measure big G and little g and get the same answers as anyone else. The quality without a name is an important concept and worth discussing, but it's not a 'scientific' concept.
For all the nagging doubts I had while reading the book and a certain amount of persistence I needed to read it to the end, there was plenty to like too. The advice on using patterns looks sound. The book says it's not enough to apply patterns by rote: it must be done with a passion. A designer must try to imagine the most wonderful space possible; must let each pattern alter the space without worrying about later patterns.
I may have quibbles with the tone of mysticism at the end of the book, but it's hardly radical to say that you have mastered a craft when you know when the rules can and should be broken. That's almost the definition of a craft. It's little wonder that the sort of people who appreciate the tradition of craft in building software appreciate TheTimelessWayOfBuilding. This book explains very well how traditional buildings have qualities that modernist buildings don't. The good ones are so much a part of their environment and so suited to their purposes that they almost become invisible. It's easy to imagine big software designs as Le Corbusier's tall buildings that satisfy formal constraints but are unnatural and inhuman. Patterns are a backlash of modesty in design and pride in workmanship. It's all very attractive to someone who works in software and values it primarily as an activity by and for people.
-- ChristianMurphy, November 2004
![[Home]](/wiki.gif)