2nd/10_Wiki/Topics/Other/Blog_Content_Rules.md

---
id: P-REINFORCE-AUTO-1FF145
title: Blog Content Rules
category: 10_Wiki/Topics
status: verified
canonical_id: self
aliases: [Blogging Rules, Content Standards]
duplicate_of: none
source_trust_level: A
confidence_score: 0.9
verification_status: applied
tags: [writing, content, blogging, style]
raw_sources: []
last_reinforced: 2026-05-10
github_commit: pending
tech_stack:
  language: markdown
  framework: hugo-astro
---

# Blog Content Rules

## 매 한 줄
> **"매 Great blog post = ONE clear thesis + concrete evidence + minimal friction"**. 2026 zero-click search era에서는 first 50 words가 entire UX를 결정 — LLM AI overview가 이미 답을 미리 보여주기 때문. 매 핵심: 매 reader's time 의 ruthlessly respect, 매 SEO theater 의 최소화.

## 매 핵심

### 매 Structural rules
- **One thesis per post**: 매 sub-thesis의 새 post로 split.
- **Inverted pyramid**: 매 conclusion-first — TL;DR 의 top.
- **Scannable hierarchy**: H2 마다 self-contained section, 매 reader 의 jump-in 가능.
- **Concrete > abstract**: 매 example의 매 claim 의 precede.

### 매 Style rules
- **Active voice default** — passive 매 deliberate choice.
- **Sentence length variance**: short. 매 medium. 매 occasional longer sentence that establishes context and rhythm before snap.
- **No hedging spirals**: "perhaps it might possibly seem that" → cut.
- **Tech terms**: define on first use, 매 jargon 의 reader-respect.

### 매 응용
1. Engineering blog post (technical deep-dive, 1500-3000 words).
2. Changelog/release notes (factual, scannable, 200-500 words).
3. Tutorial (step-by-step, runnable code, copy-paste friendly).
4. Opinion/essay (single thesis, supporting evidence, counterargument acknowledged).

## 💻 패턴

### Frontmatter template (Astro/Hugo)
```yaml
---
title: "Concrete claim, not 'Thoughts on X'"
description: "1-sentence value proposition under 160 chars for SERP"
publishDate: 2026-05-10
updatedDate: 2026-05-10
author: "Name"
tags: [primary-tag, secondary-tag]
draft: false
canonical: "https://blog.example.com/post-slug"
ogImage: "/og/post-slug.png"
---
```

### Lint rules (vale + textlint)
```yaml
# .vale.ini
StylesPath = styles
MinAlertLevel = warning

[*.md]
BasedOnStyles = Vale, write-good, Microsoft

# styles/Custom/Hedges.yml
extends: existence
message: "Hedge word '%s' — cut or commit."
level: warning
tokens:
  - perhaps
  - maybe
  - somewhat
  - quite
  - rather
  - seems to
```

### Reading-time + word-count check
```python
import re
from pathlib import Path

def analyze_post(path: Path) -> dict:
    text = path.read_text()
    body = re.sub(r"^---.*?---", "", text, count=1, flags=re.S)
    words = re.findall(r"\w+", body)
    n = len(words)
    return {
        "words": n,
        "minutes": round(n / 230),  # avg adult reading speed
        "h2_count": len(re.findall(r"^## ", body, flags=re.M)),
        "code_blocks": body.count("```") // 2,
        "links": len(re.findall(r"\[.+?\]\(.+?\)", body)),
    }
```

### TL;DR generator (Claude 4.7)
```python
from anthropic import Anthropic

def generate_tldr(post_md: str) -> str:
    client = Anthropic()
    msg = client.messages.create(
        model="claude-opus-4-7",
        max_tokens=200,
        system=("Write a 2-3 sentence TL;DR. State the thesis, the key evidence, "
                "and one actionable takeaway. No hedging. No 'this post discusses'."),
        messages=[{"role": "user", "content": post_md[:6000]}],
    )
    return msg.content[0].text
```

### SEO sanity check
```python
def seo_check(frontmatter: dict, body: str) -> list[str]:
    issues = []
    title = frontmatter.get("title", "")
    desc = frontmatter.get("description", "")
    if not (10 <= len(title) <= 60):
        issues.append(f"title length {len(title)} outside 10-60")
    if not (50 <= len(desc) <= 160):
        issues.append(f"description length {len(desc)} outside 50-160")
    if "## " not in body:
        issues.append("no H2 — flat structure hurts scannability")
    if body.count("](") < 2:
        issues.append("fewer than 2 links — orphan post")
    return issues
```

### Image optimization (sharp)
```javascript
import sharp from "sharp";

async function optimize(input, slug) {
  await sharp(input)
    .resize(1200, 630, { fit: "cover" })
    .webp({ quality: 82 })
    .toFile(`public/og/${slug}.webp`);
  await sharp(input)
    .resize(800)
    .webp({ quality: 80 })
    .toFile(`public/img/${slug}.webp`);
}
```

## 매 결정 기준
| 상황 | Approach |
|---|---|
| Tutorial | Step-numbered, runnable code, prereqs at top |
| Opinion piece | Thesis in title, counter-argument paragraph required |
| Release notes | Bulleted, version + date, breaking changes flagged |
| Long-form essay | Add ToC, 1500+ words, 3-5 H2s |
| News/timely | Publish date prominent, update date if revised |

**기본값**: 매 post 의 ship before perfect — 매 published-and-iterated 의 unpublished-and-perfect 의 superior.

## 🔗 Graph
- 변형: [[Technical Writing]]
- 응용: [[Engineering Blog]]
- Adjacent: [[SEO]]

## 🤖 LLM 활용
**언제**: TL;DR drafting, headline A/B variants, hedging-word detection, SEO description generation, outline scaffolding.
**언제 X**: full-post ghostwriting (매 voice 의 lost), factual claims requiring expertise, opinion pieces (매 author voice required).

## ❌ 안티패턴
- **SEO keyword stuffing**: 매 2026 search era에 penalize됨, reader-trust 의 erode.
- **Listicle without substance**: "10 ways to X" with shallow points.
- **Buried lede**: 매 thesis 의 paragraph 5 — modern reader 의 already gone.
- **Engagement-bait title**: "You won't believe..." — 매 trust-killer.
- **Hedging spiral**: 매 every claim 의 qualifier — reader 의 actual position 의 unclear.

## 🧪 검증 / 중복
- Verified (Strunk & White; Zinsser *On Writing Well*; Google Search Quality Rater Guidelines 2025).
- 신뢰도 A.

## 🕓 Changelog
| 날짜 | 변경 |
|---|---|
| 2026-04-20 | Auto-reinforced placeholder |
| 2026-05-10 | Manual cleanup — full substantive content, 6 patterns |