Transformational Programming Skill

Process

1. Identify the transformation boundary and steps

Express the program as input → output, then decompose it into single-responsibility pipeline steps.

Validation checkpoint: each step has one input type and one output type; no step requires knowledge of another step's internal state.

2. Implement each step as a pure-ish function

Keep transformations input-only/output-only. Use language patterns for pipeline syntax and error handling per language.

3. Compose the pipeline

Chain steps with pipe operators, method chaining, or explicit assignments:

# Explicit assignment chain (works in any language)
content = read_file(path)
lines = find_matching(content, pattern)
result = truncate(lines)
return sort_by_date(result)

4. Test each step independently

Before composing, verify each transformation works correctly with sample inputs. After composing, test the full pipeline end-to-end.

5. Handle errors

Wrap results in a success/failure type so errors propagate without breaking the chain. Each step either transforms a valid value or passes through an error unchanged. Use language patterns for language-specific patterns.

Example

User: "I need a program that reads a log file, finds ERROR entries, extracts timestamp+message, and sorts by timestamp."

## Transformation Analysis: Error Log Extractor

### Data Flow
file path → file content → error lines → parsed entries → sorted entries

### Implementation (Python)
def read_file(path):
    with open(path) as f:
        return f.read()

def filter_errors(content):
    return [line for line in content.split('\n') if 'ERROR' in line]

def parse_entries(lines):
    return [
        {'timestamp': parts[0], 'message': parts[2]}
        for line in lines
        if len(parts := line.split(' ', 2)) >= 3
    ]

def sort_by_timestamp(entries):
    return sorted(entries, key=lambda e: e['timestamp'])

def extract_errors(path):
    content = read_file(path)         # 1
    lines = filter_errors(content)     # 2
    entries = parse_entries(lines)     # 3
    return sort_by_timestamp(entries)  # 4

### Error handling
- read_file raises FileNotFoundError for missing files
- parse_entries skips malformed lines
- Steps 2 and 4 are pure and can't fail

References

• Language patterns — pipeline syntax and error handling patterns by language
• The Pragmatic Programmer (2nd ed.), Topic 30 — Transforming Programming, by Dave Thomas and Andy Hunt