Chapter Writing Standard
This guide should read like a graduate level engineering textbook or a serious technical leadership book. Its chapters should build durable mental models, show why systems take their current shape, and develop the reader's ability to reason
A rapid review of the chapter’s existing Quick Read, principles, definitions, lessons, and review material.
Purpose
This guide should read like a graduate-level engineering textbook or a serious technical leadership book. Its chapters should build durable mental models, show why systems take their current shape, and develop the reader's ability to reason under uncertainty.
Concise writing is welcome. Thin writing is not. A chapter should use prose, diagrams, examples, and evidence to develop an argument. Lists should clarify structure rather than replace explanation.
Required chapter structure
Every chapter must visibly distinguish three kinds of knowledge:
- Enduring Principle explains what should remain useful across products, vendors, models, and implementation stacks.
- Current Mission Control Implementation states only what can be traced to versioned product documents, source code, tests, and observed behavior.
- Future Vision describes desired or plausible capabilities that have not yet met the current-capability evidence bar.
The labels are epistemic boundaries. They prevent an attractive target architecture from being mistaken for working software.
Quick Read
Every new full chapter and every chapter included in a primary reading path begins with a compact Quick Read immediately after the title. It includes:
- the chapter's purpose;
- the readers who benefit most;
- prerequisite chapters or “none”;
- an honest reading-time estimate;
- the capability the reader will gain; and
- three ideas worth retaining if the reader stops there.
The Quick Read is progressive disclosure, not an abstract. It helps a reader decide whether the full chapter is relevant without pretending to replace the chapter's reasoning, evidence, tradeoffs, or exercises.
1. The problem
Describe the engineering or organizational problem before introducing a solution. Make the consequences concrete. Explain who experiences the problem and what fails when it remains unresolved.
2. Why the problem exists
Trace the incentives, constraints, historical choices, system boundaries, and failure modes that produce the problem. Avoid treating symptoms as causes.
3. The enduring principle
Introduce the principle that remains useful across vendors and technology cycles. State its scope and the conditions under which it may not apply.
4. Tradeoffs and alternatives
Explain what the principle costs, which alternatives exist, and why a different system might choose differently. Do not turn current design choices into universal laws.
5. Mission Control implementation
Show how Mission Control currently realizes the principle. Identify what is implemented, partial, proposed, experimental, or absent. Separate current fact from target architecture.
6. Future vision
Describe the desired capability only after current implementation has been bounded. State what evidence would be required to promote the vision into a current-capability claim.
7. Versioned references
Link the relevant Mission Control ADRs, product documents, code paths, tests, and evidence. Record the exact commit or version studied and the date accessed. References support the explanation; they do not replace it.
An implementation claim must cite the strongest available evidence: an ADR or accepted decision, the implementing source path, focused tests, browser or runtime behavior, and the exact commit or version. Missing evidence must be named. A claim without sufficient evidence belongs under Future Vision.
8. Notes and lessons learned
Capture the learner's own interpretation. Explain what was surprising, which assumption changed, what remains unclear, and what might be designed differently.
9. Interview and discussion questions
Include questions that test causal understanding, architecture judgment, operating-model reasoning, and the ability to distinguish vision from current implementation.
10. Whiteboard exercise
Require the learner to reconstruct the relevant boundaries, entities, flows, decisions, and failure paths without copying an existing diagram.
11. Hands-on lab
Connect the chapter to observable behavior. A lab should require code tracing, operation, implementation, debugging, validation, or recovery. It must state its prerequisites, starting version, expected evidence, and cleanup path.
Mission Control case-study rule
Mission Control is a living implementation, not the definition of an AI Software Factory. Every reference should answer:
- What principle is being studied?
- Which Mission Control decision or implementation illustrates it?
- Which commit or version was inspected?
- What tradeoff does the implementation make?
- What lesson should be retained if Mission Control later changes?
Do not copy product documentation into the guide. Link it, study it, and write the resulting understanding in original words.
Research standard
Prefer primary sources, specifications, research papers, official engineering documentation, and direct implementation evidence. Record publication and access dates. Separate vendor claims from independently established facts.
Explain conflicts between sources instead of hiding them. Mark open questions and uncertain conclusions explicitly.
Writing to avoid
Do not publish chapters that rely on slogans, unexplained jargon, repetitive bullets, unsupported forecasts, or generic AI enthusiasm. Do not present a demo, proposal, or agent assertion as proven capability.
The standard is not that every chapter must be long. The standard is that every chapter must make the reader more capable of designing, building, operating, leading, teaching, or defending an AI Software Factory.
Review this chapter.
Challenge a claim, boundary, missing failure mode, unclear term, or unsupported evidence statement.
- Claim
- Boundary
- Failure
- Evidence