Attempt of a summary of Alexander’s theory
A somewhat crude summary of Alexander’s theory could be: To create artifacts with living structure, we can’t specify the result in all detail because it is too complex. However, we can specify a generative process instead, which is much simpler to describe.
Most human artifacts are “declaratively designed” — we break everything down analytically and assemble the smallest pieces we can control. But we don’t “grow” the result. We are limited by how much detail we can understand at once, and that gives rise to abstraction, compartmentalization, and modularization, effectively ignoring parts when we think they are not important — which makes us disengage from the whole and we end up with a result that is not as coherent as it could be. The parts developed separately, work together, but they don’t really form a whole.
If we instead use the imperative process to “grow” an artifact, we divide the task differently: no longer do we compartmentalize or abstract away parts of the whole, we differentiate the part that needs most attention but keep the overall structure. The division, which is necessary for us to deal with the complexity, happens in time — we focus on a single step in a larger process of structure-preserving transformations.