Case Study Border Cleanup

Case Study: Border Window Cleanup

Type: REFACTOR Skills: cc-refactoring-guidance → cc-control-flow-quality (CHECKER) → cc-routine-and-class-design (CHECKER) Focus: Post-refactoring quality gates

Why This Case Study Matters

This example shows the CHECKER gates that run after refactoring to verify code quality wasn't degraded:

1. Two skills as quality gates - Control flow + routine design checks
2. McCabe complexity analysis - Quantified complexity with thresholds
3. Pre-existing vs new issues - Checkers distinguish what you changed from what was already there
4. Cohesion classification - Functional vs Temporal cohesion types

---

The Flow

[After refactoring work complete]

→ cc-control-flow-quality (CHECKER mode)
  → Analyzes nesting depth and McCabe complexity
  → Flags pre-existing issues separately from new code

→ cc-routine-and-class-design (CHECKER mode)
  → LSP test, parameter counts, naming, cohesion
  → Minimum viable compliance check

---

Control Flow Analysis

File	Function	Max Nesting	McCabe	Status
BorderWindow.swift	updateStyle	1 (guard clauses)	6	✓ Clean
SimpleBorderManager.swift	debugBordersImpl	2	4	✓ Clean
SimpleBorderManager.swift	setCellAssignmentsImpl	3	~14	⚠️ Pre-existing

Key insight: The checker notes setCellAssignmentsImpl complexity is pre-existing - the refactoring only added a dictionary lookup without adding control flow.

---

Routine & Class Design Check

Minimum Viable Compliance (7 checks)

#	Check	Result	Evidence
1	LSP test for inheritance	✓ PASS	No new inheritance introduced
2	Inheritance depth < 3	✓ PASS	No inheritance changes
3	No empty overrides	✓ PASS	No overrides in changes
4	Parameter count ≤ 7	✓ PASS	Max 4 params
5	Routine names describe function	✓ PASS	debugBorders, updateStyle
6	Consistent abstraction	✓ PASS	Single level
7	Implementation hidden	✓ PASS	SkyLight details encapsulated

Cohesion Classification

Routine	Type	Status
updateStyle	Functional	✓ ACCEPT - single operation
debugBorders	Functional	✓ ACCEPT - single operation
debugBordersImpl	Temporal	✓ ACCEPT - orchestration

---

CHECKER Gate Summary

Gate	Result
cc-control-flow-quality	PASS (max nesting 2, McCabe < 10 for new code)
cc-routine-and-class-design	PASS (functional cohesion, ≤7 params, good naming)

---

Key Takeaways

1. CHECKER gates after refactoring - Verify you didn't degrade quality
2. McCabe complexity - Quantified threshold (< 10 for new code)
3. Pre-existing vs new - Checkers don't blame you for existing issues
4. Cohesion types - Functional (ideal) vs Temporal (acceptable for orchestration)
5. 7-point compliance check - Systematic coverage of design principles