Doctrine

These are authoring principles, not a formal standard or certification scheme. Use them to make a professional record easier to inspect and harder to misread.

Author control

Keep an editable source record under your control. Choose what to share, with whom, and in which form. When the same claim appears in several places, maintain its meaning and evidence consistently. Repetition across websites is not independent corroboration.

Source Standard

For each material claim, name the closest available source: a test record, design revision, decision log, contemporaneous message, published artifact, or attributed account. Record its date or revision and the part that supports your statement.

Separate four kinds of statement:

• Observation: what a source records or what you directly observed.
• Recollection: what you remember, with uncertainty made explicit.
• Inference: the conclusion you draw and the reasoning that connects it to observations.
• Proposal: what you intend to do or think may work.

A source can be incomplete, mistaken, or unavailable to the reader. Say so. If you cannot substantiate a precise number, use a labeled estimate or omit it. An exact-looking number is not evidence of authenticity.

Depth Standard

A strong account explains the decision, rather than merely naming a skill. Ask:

1. What constraint made the choice difficult?
2. Which alternatives were available, and why was one selected?
3. What was your contribution, and what did others own?
4. What observation supports the outcome? Does it support causation or only a change over time?
5. Under which conditions would the conclusion stop applying?

Specificity helps a reader investigate these questions. It does not establish truth by itself. Do not use whether an AI could imitate the prose as an authenticity test.

Depth Signals

The twelve workbook techniques are prompts for describing experience. A provenance chain traces revisions; a forensic number puts a measurement in context; a decision heuristic explains a rule and its exceptions. A compressed war story makes a decision concise without hiding its limits.

A structural rhyme compares two domains. Explain both the shared mechanism and where the analogy breaks. Similar language or a shared curve does not establish an engineering equivalence. The workbook retains all twelve techniques as optional authoring aids.

Claim, evidence, project

A useful three-part authoring pattern connects:

• Claim: the bounded statement you want a reader to assess.
• Evidence: the source, observation, and limits relevant to that statement.
• Project: the context, your role, collaborators, and constraints.

This pattern is not a universal ontology or a conformance requirement. Use the structure your material needs. A private source record and a shorter public account may legitimately contain different levels of detail.

Restricted information

Separate working evidence from material approved for sharing. Keep restricted source material in an appropriate controlled environment. Review an excerpt, summary, or recollection for disclosure concerns before sharing it too; writing from memory does not remove those concerns.

Local storage is a privacy practice, not a guarantee of security or permission to disclose. When the sharing boundary is unclear, use the relevant owner’s review process before publishing the detail. This method does not determine legal compliance.

Limits and correction

Detailed accounts can still be fabricated or wrong. Sources, clear attribution, and visible uncertainty give a reviewer more to examine; they do not guarantee verification or a hiring result.

Keep a revision trail. If later evidence changes a claim, correct the source record and the public versions that repeat it. The method succeeds when the record becomes more accurate and reviewable, even when that means narrowing a claim.