Clean(er) Code for Large Scale Legacy Applications
summary
Argues that large C++ codebases (millions of LOC) can be tamed by a dedicated team, prioritized small steps, and tests; uses repo mining (Your Code Is a Crime Scene) to find hot files, approval tests and unit tests to pin behavior, profilers to find real bottlenecks, and IDEs/static analysis (Visual Assist X, ReSharper C++) plus compiler-driven refactors (extract lambda, fix compile errors) to refactor safely. Reports one year to split ~10% of UI into MVC, notes short-term slowdown but long-term faster feature delivery and fewer regressions. hot-path first compiler as refactoring aid
AI-written summary · about
chapters (10)
- 0:00Introduction to clean code
- 6:04Legacy code definition and challenges
- 16:59Why refactor large legacy codebases
- 26:44Overcoming resistance to change
- 36:09Refactoring strategies and goals
- 44:41What is refactoring and its importance
- 57:14Testing as the foundation for refactoring
- 1:06:49Incremental refactoring techniques
- 1:17:30Long-term perspective on refactoring
- 1:26:12Q&A and closing remarks
related talks
01h 01mHow to Refactor Millions of Line of Code Without Alienating your Colleagues
57mRefactoring Superpowers: Make Your C++ IDE Do Your Work, Faster and More Safely
55mSo, you inherited a large code base...
58mWho Needs C++ Unit Tests Anyway? - Modernizing Legacy Code with 0% Code Coverage
45mWorking with C++ Legacy Code
58mUnderstanding Large and Unfamiliar Codebases
01h 29mAlchemical Code Rejuvenation
01h 12mLegacy Refactoring