Designing systems we can still explain

Complexity rarely arrives announcing itself. It accumulates in imprecise names, ownerless decisions, and layers nobody can fully explain.

Explanation as a system property

An understandable system is not a simple system. It is a system where every piece has a reason that can be exposed, discussed, and checked.

«Any abstraction that cannot be formally verified is postponed technical debt.»
Systems notebook, entry 01

Complexity costC = D + V + O

A decision that can be read

Architecture is also a form of documentation. When a decision has an owner, a reason, and evidence, the system keeps its memory.

typescript

explainable-system.ts

type Decision = {
  owner: string;
  evidence: string[];
};

const explainable = (decision: Decision) =>
  Boolean(decision.owner && decision.evidence.length);

Circuit board and technical connections in dark tones — Figure 01. Architecture as a map of decisions and responsibilities.

References for further reading

The goal is not to eliminate all complexity, but to make visible which complexity is worth keeping.