Data-Driven Refactoring
summary
Argues for refactoring driven by measurable claims: validate assumptions with boundary/generative tests, approval tests and attribute checks, time the test suite and track feature-to-bug ratios, and use churn, rake notes, reek, and fukuzatsu to locate trouble spots. Describes building a simple aggregator called pandometer that ingests JSON from these tools and shows commit-by-commit trends; it takes six to eight weeks to assemble and succeeds at proving whether refactors reduce complexity and coupling over time.
AI-written summary · about
chapters (10)
- 0:00The case for refactoring
- 2:16Defining refactoring goals
- 4:33Measuring refactoring impact
- 7:56Testing strategies
- 12:25Validating assumptions
- 15:48Objective metric tracking
- 18:57Limitations of current tools
- 21:48The danger of vanity metrics
- 26:32Data-driven tool integration
- 30:06Building a custom dashboard
