Notes on "Notes on the Synthesis of Form": Dawning Insights in Early Christopher Alexander

TL;DR

This paper explores Christopher Alexander's early formalized design methods, contrasting them with his later intuitive approach, through a personal narrative and software archeology to understand his experimental computational work.

Contribution

It uncovers the historical and mathematical context of Alexander's early design synthesis methods and analyzes his experimental computer programs from the 1950s and 1960s.

Findings

Alexander's early work involved formalized, modular design methods.

His computer programs aimed to implement his design principles.

Replication attempts revealed challenges and limitations in his approach.

Abstract

This essay is a picaresque -- a first-person narrative relating the adventures of a rogue (me) sifting through the mind of Christopher Alexander as he left behind formalized design thinking in favor of a more intuitive, almost spiritual process. The work of Christopher Alexander is familiar to many computer scientists: for some it's patterns, for some it's the mystical **quality without a name** and "Nature of Order"; for many more it's "Notes on the Synthesis of Form" -- Alexander's formalized design method and foreshadowing ideas about cohesion and coupling in software. Since the publication of "Design Patterns" by Gamma et al. in 1994, there have been hundreds of books published about design / software patterns, thousands of published pattern languages, and tens of thousands of published patterns. "Notes," published in 1964, was quickly followed by one of Alexander's most important essays, "A City is Not a Tree," in which he repudiates the formal method described in "Notes," and his Preface to the paperback edition of "Notes" in 1971 repeats the repudiation. For many close readers of Alexander, this discontinuity is startling and unexplained. When I finally read "Notes" in 2015, I was struck by the detailed worked example, along with a peculiar mathematical treatment of the method, and a hint that the modularization presented in the example was reckoned by a computer program he had written -- all in the late 1950s and early 1960s. Because of my fascination with metaheuristic optimization, I couldn't resist trying to replicate his experimental results. Computers and their programs relish dwelling on flaws in your thinking -- Alexander was not exempt. By engaging in hermeneutics and software archeology, I was able to uncover / discover the trajectory of his thinking as he encountered failures and setbacks with his computer programs. My attempted replication also failed, and that led me to try to unearth the five different programs he wrote, understand them, and figure out how one led to the next. They are not described in published papers, only in internal reports. My search for these reports led to their being made available on the Net. What I found in my voyage were the early parts of a chain of thought that started with cybernetics, mathematics, and a plain-spoken computer; passed through "A City is Not a Tree"; paused to "make God appear in the middle of a field"; and ended with this fundamental design goal: **I try to make the volume of the building so that it carries in it all feeling. To reach this feeling, I try to make the building so that it carries my eternal sadness. It comes, as nearly as I can in a building, to the point of tears.**