Foreword
Throughout much of human history, our technological achievement was driven by the quest for certainty and precision. Chaos in all its forms had to be defeated by the clockwork-like precision of the machines we built. We altered the course of flow of rivers. We achieved space flight. We created hierarchical societal structures with protocols and processes. But despite the mathematical precision, we never managed to achieve the robustness and efficiency of natural processes.
Our human intuition tells us that precise understanding of the underlying processes and consistency of their execution are key to well-functioning systems. This belief gives us a seemingly practical mental framework where all conditions affecting such systems are known and understood, or, at least, can be approximated or anticipated in some way. We further quantize the behaviors into sets of well-understood rules and actions and combine them in hierarchically organized global knowledge. A system of consistency hierarchies.
This approach led us to industrialization, but it also gave us an equivalent of industrialized micromanagement. Our control and automation systems turned into the equivalent of sequenced linear execution plans, where each step is a precise action—certain, predictable, and consistent. Any significant inconsistency leads to local and often cascading failure, at times, with global consequences—all with little or no ability to reconverge.
Our cultural preference for understanding the complete picture led us to the creation of control systems that function as large, centralized, all-knowing brains that make precise decisions based on assumptions of availability and consistency, resulting in actions that are elemental and imperative in nature—our intuitive strive to micromanage the reality. Such centralization, even if it’s logical, leads to inadvertent limitations in scale. A centralized control brain, as large as it can be, still has finite processing capacity. A busy brain can’t make decisions fast enough to carry out micromanaging actions in time. Latency of actions creates imprecision and inconsistency, often with unknown consequences.
In this book, Mark Burgess discusses a different philosophical approach to solving large-scale control problems that entirely eliminates the need for a big almighty brain and the very need for micromanagement. Instead, he focuses on breaking the system down into a large number of primitive functions, each fully autonomous, self-enforced, with the ability to deal with imprecision at the lowest level possible. Like in nature, these functions are not centrally orchestrated—they act like organisms that interact with each other in the form of promises, not subjected to absolute universal obligations, not requiring precise consistent knowledge.