Code Example Best Practices

Guidelines for writing clear, effective code examples in technical blog posts.

Overview

Code examples are often the most valuable part of technical content. This skill provides standards for writing code snippets that are easy to understand, copy, and adapt.

This skill covers:

• Code snippet formatting and structure
• Comment and annotation guidelines
• Context and setup requirements
• Error handling in examples

This skill does NOT cover:

• Prose writing style (see technical-writing-style)
• Overall post structure (see content-structure-patterns)

Quick Reference

Code Block Essentials

Element	Guideline
Language tag	Always specify (python, bash, etc.)
Length	Under 30 lines preferred
Width	Under 80 characters per line
Comments	Explain "why", not "what"
Imports	Include when relevant to example

Example Quality Checklist

• Runs without modification
• Language tag specified
• Non-obvious lines commented
• Variables have meaningful names
• Sensitive values use placeholders
• Expected output shown (where helpful)

Principles

1. Make Examples Runnable

Code that doesn't run frustrates readers. Every example should:

• Include necessary imports
• Define required variables
• Use realistic (but safe) placeholder values
• Work in a standard environment

Good:

import os
from pathlib import Path

# Read config from environment (provide default for local dev)
api_key = os.environ.get("API_KEY", "your-api-key-here")
config_path = Path("config.json")

if config_path.exists():
    config = json.loads(config_path.read_text())

Bad:

# Missing imports, undefined variables
config = json.loads(config_path.read_text())

2. Keep Examples Focused

Show only what's necessary. Strip everything else.

Good:

# Retry with exponential backoff
for attempt in range(max_retries):
    try:
        return make_request(url)
    except RequestError:
        sleep(2 ** attempt)
raise MaxRetriesExceeded()

Bad:

# Too much context obscures the pattern
import requests
import logging
from typing import Optional
from dataclasses import dataclass

logger = logging.getLogger(__name__)

@dataclass
class Config:
    max_retries: int = 3
    base_url: str = "https://api.example.com"
    timeout: int = 30

def make_request(url: str, config: Optional[Config] = None) -> dict:
    config = config or Config()
    # ... 20 more lines before getting to the retry logic

3. Use Progressive Disclosure

For complex examples, build up incrementally:

Step 1: Basic version

def process_data(data):
    return [item.upper() for item in data]

Step 2: Add error handling

def process_data(data):
    results = []
    for item in data:
        try:
            results.append(item.upper())
        except AttributeError:
            results.append(str(item).upper())
    return results

Step 3: Add logging (optional)

def process_data(data, logger=None):
    results = []
    for item in data:
        try:
            results.append(item.upper())
        except AttributeError:
            if logger:
                logger.warning(f"Converting {type(item)} to string")
            results.append(str(item).upper())
    return results

4. Show Expected Output

Help readers verify they're on track:

$ curl -s https://api.example.com/health | jq
{
  "status": "healthy",
  "version": "1.2.3"
}

>>> calculate_hash("hello world")
'b94d27b9934d3e08a52e52d7da7dabfac484efe37a5380ee9088f7ace2efcde9'

Formatting Standards

Language Tags

Always specify the language for syntax highlighting:

```python
def example():
    pass
```

Common tags: python, javascript, typescript, bash, json, yaml, sql, go, rust

Line Length

Keep lines under 80 characters to prevent horizontal scrolling:

# Good: Line breaks at logical points
result = (
    some_long_function_name(
        parameter_one=value,
        parameter_two=other_value,
    )
)

# Avoid: Long lines that scroll
result = some_long_function_name(parameter_one=value, parameter_two=other_value, parameter_three=another_value)

Comments

Comment the "why", not the "what":

# Good: Explains reasoning
# Use a set for O(1) lookup on large datasets
seen = set()

# Bad: States the obvious
# Create a set called seen
seen = set()

Inline comments for non-obvious lines:

response = client.get(url, timeout=30)  # Server can be slow during peak hours
data = response.json().get("results", [])  # API returns empty list as null

Placeholder Values

Use obvious placeholders that won't accidentally work:

Type	Good Placeholder	Bad Placeholder
API keys	your-api-key-here	abc123
URLs	https://api.example.com	https://api.com
Passwords	<your-password>	password123
Emails	user@example.com	test@test.com
IDs	12345 or <user-id>	1

Diffs and Changes

Show what changed when modifying code:

 def process_data(data):
-    return data.upper()
+    return data.strip().upper()

Or use comments to highlight changes:

def process_data(data):
    return data.strip().upper()  # Added strip() to handle whitespace

Common Patterns

Configuration Examples

Show both environment variables and code:

# Set in your shell or .env file
export DATABASE_URL="postgres://user:pass@localhost:5432/mydb"
export REDIS_URL="redis://localhost:6379"

import os

DATABASE_URL = os.environ["DATABASE_URL"]
REDIS_URL = os.environ.get("REDIS_URL", "redis://localhost:6379")

Command Line Examples

Use $ prefix for commands, show output without prefix:

$ npm install express
added 57 packages in 2.3s

$ npm start
Server running on http://localhost:3000

Multi-File Examples

When showing multiple files, use clear headers:

src/config.py

DATABASE_URL = "postgres://localhost/mydb"

src/main.py

from config import DATABASE_URL

Error Examples

When showing errors, include enough context to diagnose:

>>> import missing_module
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ModuleNotFoundError: No module named 'missing_module'

Anti-Patterns

Screenshot of Code

Never use images for code. Always use text that readers can copy.

Untested Examples

Run every example before publishing. Typos and API changes break tutorials.

Missing Context

Don't assume readers know the file structure or have run previous steps:

# Bad: Where does 'client' come from?
response = client.get("/users")

# Good: Show the setup
from myapp import create_client
client = create_client(api_key=os.environ["API_KEY"])
response = client.get("/users")

Hardcoded Secrets

Never include real credentials, even in "example" form:

# NEVER do this
api_key = "sk-live-abc123..."  # This looks like a real key

# Do this instead
api_key = os.environ["API_KEY"]

See Also

• technical-writing-style skill - Writing prose around code
• content-structure-patterns skill - Where code fits in post structure