Writing Standards

Voice

How we write across the CREATE SOMETHING ecosystem. Not a style guide—a philosophical commitment to clarity, honesty, and empirical rigor.

Five Principles

1. Clarity Over Cleverness

Code is read 10x more than written. Writing is read 100x more than written. Serve the reader.

Yes: "Built in 6 hours vs 15-20 hours estimated (60-70% savings)"

No: "Significantly improved development velocity"

2. Specificity Over Generality

Every claim must be measurable. Every metric must be precise. Vagueness is dishonest.

Yes: "26 hours actual vs 120 estimated"

No: "Saved significant development time"

3. Honesty Over Polish

Document failures. Acknowledge limitations. Share what didn't work as freely as what did.

Required: "What This Proves / What This Doesn't Prove" sections

Required: "Where User Intervention Was Needed" documentation

4. Useful Over Interesting

Every paper must answer: Is this useful? Not interesting, not novel—useful. Can readers implement this? Are prerequisites clear?

Required: Reproducibility sections with starting prompts

Required: Expected challenges documented

5. Grounded Over Trendy

Connect technical decisions to timeless principles. Cite masters. Show philosophical lineage. Principles that withstand technological change.

Example: "Given: Rams 'Good design is unobtrusive'"

Example: "Context: Tufte's principle of maximizing data-ink ratio"

Sentence Patterns

Short Declarative Statements

"This isn't minimalism for aesthetics. It's discipline for clarity."

"Less, but better."

"Decoration is dishonest."

Paired Constructions (What It Is / Isn't)

"Not blog posts about AI, but real data from building real systems"

"Not features for features' sake"

"Not interesting, not novel—useful"

Em-Dash for Emphasis

"Built in 6 hours — 65% faster than manual development"

"These aren't experiments — they're business-critical infrastructure"

"Research papers with tracked experiments — not just blog posts"

Interrogative (Questions That Interrogate)

"Is my design good design?"

"Does this meet Rams' principles?"

"What must remain?"

Required Elements for Experiments

ASCII Art Header — Terminal aesthetic with key metrics displayed
Hypothesis → Validation — State what you're testing, measure results
Success Criteria — Checkbox format with ✅/❌ actual outcomes
Metrics Table — Time, cost, savings, ROI calculated
Honest Assessment — "What This Proves" AND "What This Doesn't Prove"
Where Intervention Was Needed — Document errors, iterations, learnings
Reproducibility — Prerequisites, starting prompt, expected challenges
Master Citation — Connect to canonical principles when relevant
Hypothesis Outcome — ✅ VALIDATED or ❌ INVALIDATED (be honest)

Forbidden Patterns

Marketing Jargon (Never Use)

Cutting-edge
Revolutionary
Game-changing
AI-powered
Leverage
Synergy
Solutions
Best-in-class

Vague Claims (Replace With Specifics)

"Significantly improved performance"

→ "Reduced load time from 3.2s to 0.8s (75% improvement)"

"Many users benefited"

→ "47 active users across 12 organizations"

Decoration Without Function

No emoji (unless terminal aesthetic like ✅/❌ for status)

No stock photos for visual interest

No color accents for "brand personality"

Every element must justify its existence

Voice Checklist

Before publishing any content, verify:

[ ] Are all claims backed by specific metrics?
[ ] Are time/cost comparisons included?
[ ] Is the methodology transparent?
[ ] Are failures documented?
[ ] Are limitations acknowledged?
[ ] Is this reproducible by others?
[ ] Does this solve a real problem?
[ ] Can readers implement this?
[ ] Are prerequisites clear?
[ ] Is a master/principle cited when relevant?
[ ] Does this connect to the canon?
[ ] No marketing jargon?
[ ] Direct, declarative sentences?
[ ] Specific over vague?

Preferred Terminology

AI Development

Use: "AI-native development" (not "AI-assisted" or "AI-powered")

Use: "Agentic systems" (not "AI agents" alone)

Use: "Claude Code" (specific tool, not generic "AI")

Use: "Working with AI agents as development partners" (not "using AI")

Research

Use: "Experiments" (not "projects" or "case studies")

Use: "Papers" (not "blog posts" or "articles")

Use: "Tracked experiments" (emphasis on measurement)

Use: "Reproducible results" (not "findings")

Use: "Rigorous methodology" (not "best practices")

Quality

Use: "Production-ready" (not "functional" or "working")

Use: "Business-critical infrastructure" (not "enterprise solution")

Use: "Systems thinking" (not "best practices")

Use: "Canonical standards" (not "style guide")

Philosophy

Use: "Less, but better" (always this phrase)

Use: "Weniger, aber besser" (German original for formal contexts)

Use: "Modes of Being" (not "properties" or "websites")

Use: "The Canon" (capitalized, sacred)

Use: "Masters" (not "influences" or "inspiration")

The Hermeneutic Test

Every piece of content must pass the hermeneutic circle. This is our quality control:

Does this part reveal the whole?

Does this experiment embody "less, but better"? Can someone read this and understand what CREATE SOMETHING stands for?

Does the whole explain this part?

Can you trace this decision to a canonical principle? Does this connect to .ltd standards?

Does this strengthen the circle?

Does it feed back to validate or evolve the canon? Does it make the ecosystem more coherent?

If any answer is "no," revise or reject. The circle is our competitive advantage.

How the Masters Wrote

Our voice is isomorphic to our principles—the form matches the content. This is how the masters wrote:

Dieter Rams

Declarative principles. No fluff. Compressed wisdom.

"Good design is as little design as possible."

Ludwig Mies van der Rohe

Aphoristic. Architectural. Essential.

"Less is more."

"God is in the details."

Edward Tufte

Empirical descriptions paired with visual proof. High data density.

"Above all else show the data."

Charles and Ray Eames

Functional elegance. Every word serves the reader.

"The best for the most for the least."

We don't imitate their style. We enact the same discipline that produced their style. Writing that emerges from "weniger, aber besser" sounds like this because it must.

This is not a brand guideline.

Brand guidelines are static rules. This is a living interpretive framework. The voice evolves through the hermeneutic circle—slowly, carefully, through validation in real-world practice.

Protect the circle. Trust the circle. Let the circle refine itself.