Frontend: Graphs. Build. Dead Code
summary
Nikita from Yandex.Market explains using a dependency graph extracted from the build to find and remove dead code in a large frontend repository. He describes enriching the build graph with additional sources (TODOs, experiments, API usage) because the raw dependency graph alone is insufficient, and turning deletion into a reproducible process and development culture. He also presents metrics showing benefits and experiments with network-analysis tools to assess repository health.
AI-written summary · about
chapters (29)
- 0:00Welcome
- 4:54Talk goal
- 6:22Yandex Market architecture
- 7:26Code growth and dead code
- 9:13Dead code definition
- 11:44Finding dead code
- 12:46Build dependency graph
- 15:12Dependency analysis tools
- 17:17Base build graph
- 18:21Build optimizations impact
- 20:51Re-exports problems
- 22:58Code removal algorithm
- 25:23Module filtering and tests
- 27:56Cross-imports problems
- 29:42Docs and tests barriers
- 31:06Logs analysis and metrics
- 32:50Removing features
- 35:57Usage metrics
- 37:42Removal results
- 39:27Dependency analysis rules
- 41:55Graphs as networks
- 43:20Yandex Market graph visualization
- 45:45Graph metrics: degree and clustering
- 47:30Degree distribution
- 49:59Centrality and importance
- 52:28Modularity and extraction
- 54:57Module authority metric
- 56:20Analysis granularity
- 57:46Q&A
related talks
30mAutomatic Visualizations of the Frontend
28mI've Fallen Into a Large Codebase and I Can't Get Up!
48mFrom Chaos to Clarity Mathematical Foundations for Maintainable React Applications
40mA study of a large-scale micro component architecture
44mBuilding resilient frontend architecture
18mUnderstand Your Codebase to Innovate Faster
40mBuilding resilient frontend architecture
41mGraphs, not trees: A ground-up approach to fixing a docs site.