What is your Algorithmic Core?
summary
Production C++ often hides small but critical algorithms; this talk shows how to extract and test them by example: a linear-time substring matcher implementing KMP LPS array passes 17 naive tests, then still fails until mutation testing and an aba fractal corner case are added; a cached recursive dependency routine silently misorders levels until replaced by an explicit topological-sort graph; a nested subcase test runner exposes a hidden rooted-tree execution algorithm. LLMs help generate and refactor tests but sometimes hallucinate.
AI-written summary · about
chapters (7)
- 0:00Algorithmic core concepts
- 7:25Engineering vs algorithm bugs
- 13:41Why standard testing fails
- 22:11Hidden topological sorting
- 24:59Algorithm complexity in tests
- 30:40Identifying algorithmic smells
- 39:14Q&A on LLM bugs
related talks
59mTheory makes beautiful programs
47mFantastic Algorithms and Where To Find Them
01h 01mAlgorithm Intuition (part 2 of 2)
01h 04mBetter Code: Data Structures
56mIdentifying Common Code Smells
56mWill Your C++ Program Still Be Correct Next Year?
01h 30mTuning C++: Benchmarks, and CPUs, and Compilers! Oh My!
01h 35mThe Twin Algorithms