A Structured Approach to Problem Solving and Architectural Thinking

This article reflects on the concepts presented in "About Architecture – Part One: What Is a Problem" and expands them into a practical framework for analyzing and solving both anomalous and improvement‑type problems.

Problem Definition

Problems are defined as a gap between a goal and the current state. They can be classified as anomalous (unexpected failures) or improvement (desired enhancements). The article emphasizes that architecture exists to translate real‑world requirements into software solutions, and that solving problems requires moving from an open‑ended exploration to a closed, concrete implementation.

Common Pitfalls

Three typical mistakes are illustrated with concrete examples: (1) reacting to an alert without root‑cause analysis, (2) demanding a solution from others without first understanding the underlying problem, and (3) executing a task without clarifying its ultimate purpose. Each case is followed by an analysis that stresses the need for deep inquiry.

The Learning Golden Ring

The author proposes a "what‑how‑why" learning loop: first ask what something is, then how it works, and finally why it is designed that way. This loop drives deeper understanding and fuels continuous improvement.

Solving Anomalous Problems

The key is to locate the root cause (often called the "who" of the problem). A popular technique such as the 5W1H analysis is mentioned. The article describes two ways to model problem stakeholders: a linear chain and a tree structure. An example of a web page outage demonstrates how to trace the chain from the user’s device, network, CDN, load balancer, to the backend service, and then examine each link for failures.

Solving Improvement Problems

Improvement problems start with a clear future goal. The process is broken into three steps: what (define the core objective), who (identify all stakeholders, their responsibilities, and interests), and how (allocate limited resources to satisfy the most critical stakeholder needs). The article provides a detailed enumeration of stakeholders in a fictional scenario involving a boss, a senior official, a secretary, a driver, and an employee named Xiao Li.

Human Factors

The author notes that most problems involve people, and effective problem‑solvers need high emotional intelligence. References to books such as "How to Win Friends and Influence People" and "Thinking, Fast and Slow" are offered for further reading.

Summary

Learning Golden Ring: a continuous what‑how‑why loop.

Anomalous problems: locate the "who" by mapping stakeholder chains (linear or tree).

Improvement problems: clarify the core goal, enumerate stakeholders, and prioritize key interests.

Human nature: understanding and sometimes counteracting human behavior is essential for solving people‑related issues.

The article concludes by hinting at a forthcoming third part that will delve deeper into software architecture.