It's a small world
summary
Claims are concrete not vague: Watts and Strogatz define small‑world networks with high clustering and low path length, but module graphs in Elixir turn out treelike. A custom parser parses Elixir AST and exports dependency graphs to a JVM graph UI, then 14 apps are measured (5 Phoenix, 9 libraries) and clustering coefficients sit around ~0.2–0.33 with Hexpm showing ~350 modules and >250 nodes at zero. Conclusion: code decomposes into trees, clustering flags missing abstractions; the parser and results live on GitHub.
AI-written summary · about
chapters (4)
- 0:00Start
- 5:34CALCULATING A CLUSTERING COEFFICIENT
- 8:07CALCULATING AVERAGE SHORTEST PATH
- 24:42FURTHER RESEARCH IDEAS
related talks
30mSolid Modularization - Untangling the dependency graph
48mFrom Chaos to Clarity Mathematical Foundations for Maintainable React Applications
40mLightning Talks
25mDetangling Your Dependency Graph
38mThree Games In Elixir
01h 00mFrontend: Graphs. Build. Dead Code
22mYou Don't Know Your Dependencies
38mCode is the Easy Part