Technical Writing

Produces technical documents that apply progressive disclosure, precise terminology, and rigorous evidence attribution to make complex engineering content accessible, navigable, and trustworthy.

Guiding Principle

"Technical writing is not about demonstrating expertise. It is about transferring understanding efficiently."

Procedure

Step 1 — Audience and Purpose Analysis

1. Define the primary and secondary audiences with their technical depth.
2. Establish the document's purpose: inform, persuade, instruct, or reference.
3. Determine the reading context: deep study, quick reference, or executive scan.
4. Select the appropriate document structure and depth level.

Step 2 — Content Architecture

1. Design the information hierarchy using progressive disclosure (TL;DR first).
2. Structure sections to support both linear reading and random access.
3. Define the terminology glossary for domain-specific or ambiguous terms.
4. Plan visual elements: diagrams, tables, and code samples with placement rationale.

Step 3 — Drafting with Evidence

1. Write each section following the claim-evidence-implication pattern.
2. Tag every assertion with its evidence type: [CODE], [CONFIG], [DOC], [INFERENCE], [ASSUMPTION].
3. Use precise language: avoid hedging words unless uncertainty is genuine.
4. Include cross-references to related sections and external sources.

Step 4 — Quality Review

1. Verify terminology consistency throughout the document.
2. Check that progressive disclosure works: each level is self-contained.
3. Validate all evidence tags and cross-references.
4. Test readability against the target audience profile.

Quality Criteria

• Document opens with a self-contained TL;DR that serves executives and scanners.
• Every technical assertion has an evidence tag.
• Terminology is consistent and defined at first use.
• Progressive disclosure enables reading at three depth levels.

Anti-Patterns

• Writing for peers instead of the defined audience.
• Front-loading methodology descriptions before delivering findings.
• Using evidence tags inconsistently or omitting them for inferred conclusions.
• Creating documents that require sequential reading with no random access support.