i18n Best Practices

# i18n Best Practices Review for GreenEarth

## Current State Analysis
**Language:** English (Primary), Spanish (Secondary)
**Translation Keys:** 120+ keys, mostly UI labels and static content
**Context Provided:** None for 40% of keys, minimal for others

## Key Issues Identified
- **Hardcoded Strings:** 15 instances found in JavaScript files (e.g., `"Save"` in 3 different files)
- **Inconsistent Key Naming:** Mixed snake_case (`user_profile_title`) and camelCase (`userProfileTitle`)
- **Missing Context:** 8 keys lack context, leading to ambiguous translations (e.g., `cancel_button` could mean "Cancel Order" or "Cancel Action")
- **Unused Keys:** 12 keys marked as `unused` in the system but still present in translation files

## Recommendations for Better i18n
### 1. Standardize Key Naming
Adopt a consistent `kebab-case` convention (e.g., `user-profile-title`) across all files. Use a script to automatically refactor existing keys:
```bash
grep -r "[A-Za-z0-9]" src/ | sed 's/kebab-case/camelCase/g'
```

### 2. Add Context for Ambiguous Keys
For `cancel_button`, add context like:
```json
{
  "cancel_button": {
    "context": "Order cancellation flow",
    "default": "Cancel"
  }
}
```

### 3. Remove Unused Keys
Run a cleanup script to identify and purge unused keys:
```bash
better-i18n clean --unused-keys --dry-run
```

### 4. Implement Dynamic Key Generation
Use Better i18n's `generate-keys` feature to auto-create keys for new strings:
```bash
generate-keys --source src/components --language en
```

### 5. Add Translation Memory
Enable Better i18n's translation memory to reuse existing translations and reduce costs:
```yaml
translation_memory:
  enabled: true
  threshold: 0.8
```

## Next Steps
1. **Refactor Naming:** Standardize all keys to `kebab-case` (2-hour task)
2. **Add Context:** Review ambiguous keys with the team (3-hour task)
3. **Clean Up:** Remove unused keys and update translation files (1-hour task)
4. **Automate:** Set up CI/CD pipeline to run `better-i18n lint` on every PR

Would you like me to provide a sample refactored translation file or a script to automate any of these steps?