Do Physical Laws Exist in the Universe of Code?
Cross-analysis of 29 repositories, 55,343 engineers, and 5 language families
EIS (Engineering Impact Signal) is a CLI tool that quantifies the "structural influence" an engineer has left on a codebase, using only data from git log and git blame.
In this analysis, we use Gravity Concentration (what percentage of total Gravity the top 10 engineers hold) as the cross-language comparison metric.
Different languages produce different "physical laws." Type system expressiveness, framework culture, and architectural concentration.
We classified 29 OSS repositories into 5 categories and compared the distribution of structural authority.
| Category | Characteristics | Languages | Repos |
|---|---|---|---|
| Expressive | Rich type systems, ADTs, pattern matching, traits | Rust, Scala | 5 |
| Go (Self-structured) | Static typing, anti-framework culture, explicit interfaces | Go | 7 |
| Framework-driven | Structure delegated to frameworks, CoC | Ruby, PHP, Java, Python, TS, Elixir | 6 |
| Systems (C/C++) | Manual memory management, templates | C, C++ | 5 |
| Dynamic / Structural | Dynamic typing, self-structured | JavaScript, TypeScript, Python | 6 |
What percentage of total Gravity is held by the top 10 engineers. Higher values mean "structure is concentrated in fewer architects."
Structure is built by "people." A few architects hold overwhelming design authority. Concentration: High.
Structure is embedded in "types." Design authority is distributed across everyone who writes type signatures. Concentration: Low.
Structure is absorbed by the framework. However, internal governance spans a spectrum from distributed to concentrated.
Both belong to the "Framework-driven" category. Both are popular frameworks with legendary creators.
Yet their Gravity physics are completely opposite.
125 contributors exist, but one person holds 92.5% of all structural gravity.
React, with 2,010 engineers involved, has undergone 5 generations of architect transitions over 10 years.
Using EIS's timeline feature to track Gravity in 3-month spans, the moments of generational transition become visible.
EIS is computed solely from git log and git blame. It cannot see RFCs or design documents.
Yet the results are surprisingly consistent with community intuition.
| Validation Item | EIS Result | Community Intuition | Match |
|---|---|---|---|
| Rails design distribution | Top10 Design 51.2, 6 architects | "Rails has a deep core team" | Match |
| Laravel design concentration | Taylor Otwell = Design 100, all others < 4 | "Laravel is Taylor's creation" | Match |
| esbuild's solo nature | 92.5% concentration, Evan Wallace all-axis 100 | "esbuild was built by Evan alone" | Match |
| Kubernetes distribution | Gravity Conc. 0.8% | "K8s has a massive community" | Match |
| React's generational shifts | 5 Anchor transitions over 10 years | "React is constantly evolving" | Match |
| Jorge Cabiedes | Design 100 with 82 commits | Embodiment of "quality over quantity" | Match |
| Go vs Rust concentration | Go 16.4% vs Rust/Scala 6.7% | "Go relies on few architects for design" | Match |
git blame, it reflects the structural truth of a project with surprising accuracy.Gravity concentration varies by 4.8x across language families — from Go (16.4%) to Systems C/C++ (3.4%).
Three structural authority modes — Architect-centric (Go), Type-distributed (Rust/Scala), Framework-absorbed (Rails/Laravel) — show that even when code looks similar, the physics of structural authority are fundamentally different.
Even within the same Framework-driven category, Rails and Laravel exhibit opposite governance physics. esbuild is a singularity where physical laws break down. React achieved healthy generational transition over 10 years.
Knowing which physical laws govern your universe is the first step toward better structural decisions.
| Repository | Category | Language | Engineers | Top10 Design | Top10 Survival | Grav Conc |
|---|---|---|---|---|---|---|
| esbuild | Go | Go | 125 | 10.0 | 10.0 | 92.5% |
| express | Dynamic | JS | 391 | 13.5 | 3.2 | 23.9% |
| swc | Expressive | Rust | 357 | 16.4 | 14.0 | 15.7% |
| prettier | Dynamic | JS | 799 | 42.9 | 10.0 | 13.4% |
| fastapi | Framework | Python | 897 | 11.7 | 11.2 | 9.4% |
| vite | Dynamic | TS | 1,247 | 32.6 | 16.0 | 9.3% |
| eslint | Dynamic | JS | 1,180 | 18.6 | 12.8 | 8.1% |
| loki | Go | Go | 1,319 | 34.0 | 19.3 | 7.8% |
| spring-boot | Framework | Java | 1,504 | 30.2 | 34.1 | 7.3% |
| polars | Expressive | Rust | 694 | 17.4 | 29.8 | 7.0% |
| nest | Framework | TS | 684 | 13.3 | 10.2 | 7.0% |
| scala | Expressive | Scala | 794 | 32.3 | 5.9 | 6.8% |
| prometheus | Go | Go | 1,281 | 44.5 | 11.9 | 5.4% |
| superset | Dynamic | Python/TS | 1,477 | 23.6 | 29.5 | 4.5% |
| duckdb | Systems | C++ | 690 | 20.7 | 4.9 | 4.5% |
| phoenix | Framework | Elixir | 1,424 | 18.9 | 14.4 | 4.5% |
| redis | Systems | C | 905 | 14.3 | 8.9 | 4.0% |
| argo-cd | Go | Go | 1,889 | 41.4 | 20.5 | 4.0% |
| arrow | Systems | C++ | 1,466 | 28.3 | 12.4 | 3.8% |
| envoy | Systems | C++ | 1,444 | 46.4 | 15.7 | 3.5% |
| react | Dynamic | JS | 2,010 | 17.5 | 31.1 | 3.4% |
| scala3 | Expressive | Scala 3 | 995 | 12.6 | 12.9 | 3.3% |
| terraform | Go | Go | 2,223 | 30.3 | 32.7 | 2.4% |
| grafana | Go | Go/TS | 2,893 | 16.3 | 24.6 | 1.9% |
| laravel | Framework | PHP | 4,449 | 17.2 | 15.3 | 1.5% |
| ClickHouse | Systems | C++ | 2,563 | 17.3 | 30.0 | 1.4% |
| rails | Framework | Ruby | 6,512 | 51.2 | 7.3 | 0.9% |
| kubernetes | Go | Go | 5,217 | 42.5 | 19.9 | 0.8% |
| rust | Expressive | Rust | 7,914 | 53.0 | 5.7 | 0.6% |