Categories of the Commons: Formalizing Open Source Governance with Mathematics

Why do some open source projects thrive for decades while others collapse under their own success? Why did Node.js fork into io.js? Why do single-maintainer projects like curl sustain critical infrastructure while well-funded foundations sometimes struggle? I’m announcing Categories of the Commons, my MBA thesis research that applies category theory, sheaf cohomology, and cybernetics to formalize open source governance—and potentially predict governance crises before they happen.

The Problem

Open source software powers everything. Your browser, your phone, the servers behind every website—all built on a foundation of commons-based peer production. Yet we lack formal frameworks for understanding why some projects succeed while others fail.

Consider these patterns:

curl has been maintained primarily by Daniel Stenberg for 25+ years, serving as critical infrastructure for millions of systems
core-js sustains the JavaScript ecosystem while its maintainer struggles financially
Kubernetes thrives with hundreds of contributors and formal governance
io.js forked from Node.js due to governance conflicts, then merged back

What makes these outcomes different? Can we formalize the difference?

The Framework

My research synthesizes four intellectual traditions:

1. Stafford Beer’s Viable System Model (VSM)

Beer identified five systems necessary for organizational viability. In OSS terms:

System	Function	OSS Manifestation
S1	Operations	Commits, PRs, releases
S2	Coordination	CI/CD, review processes, CONTRIBUTING.md
S3	Control	Metrics, release management
S4	Intelligence	Roadmaps, ecosystem monitoring
S5	Policy	Governance docs, mission, identity

2. Elinor Ostrom’s Commons Governance

Ostrom won the Nobel Prize for showing that commons can be governed successfully. Her 8 design principles map to OSS—but with a twist: they predict success for Federations (Kubernetes, Linux) but fail for Stadiums (curl, SQLite).

3. Nadia Asparouhova’s OSS Taxonomy

From Working in Public, four project types based on user growth vs. contributor growth:

                    CONTRIBUTOR GROWTH
                    Low              High
                ┌─────────────┬─────────────┐
           High │   STADIUM   │ FEDERATION  │
  USER          │  curl, npm  │  Linux, K8s │
 GROWTH         ├─────────────┼─────────────┤
           Low  │    TOYS     │    CLUBS    │
                │  Personal   │  Niche FWs  │
                └─────────────┴─────────────┘

4. Category Theory & Sheaf Cohomology

Here’s where it gets interesting. Category theory provides the compositional semantics—how governance structures combine and transform. But the real innovation is applying sheaf theory:

Treat OSS projects as topological spaces
Governance rules become sheaf sections
The gluing axiom captures how local decisions must combine into coherent global policy
Čech cohomology measures governance coherence

The cohomology groups have concrete interpretations:

Group	Meaning
H⁰	Global consensus (universal rules)
H¹	Governance conflicts (incompatible local policies)
H²	Structural obstructions (deep incompatibilities)

The Hypothesis

Main Conjecture: Non-trivial H² cohomology classes precede fork events by 6-12 months.

In plain terms: when a project has deep structural governance incompatibilities—situations where three parties A, B, C agree pairwise but can’t all agree together—a fork becomes likely.

This is testable. We can reconstruct historical project states for known forks (Node.js/io.js, Bitcoin/Bitcoin Cash, OpenOffice/LibreOffice) and check if H² spiked before the split.

The Research Design

I’m taking a Stadium-focused approach:

28-30 Stadium projects (curl, core-js, axios, etc.) — maximum categorical signal
12-15 Federation projects (Kubernetes, Rust, Node.js) — baseline
8-10 Club projects — convergent case
15-20 Control projects — noise estimation

Why Stadium-heavy? These projects are terminal objects in the organizational constraint category. With ≤3 maintainers handling massive usage, they exhibit the clearest governance structure-to-entropy mapping.

Early Findings

I’ve collected data on 13 projects so far. Early patterns emerging:

Entropy correlates with governance type — Stadium projects show distinct entropy profiles
Dominance ratio > 40% is a strong Stadium indicator
Governance files matter — presence of CONTRIBUTING.md, CODEOWNERS significantly affects cohomology

What’s Next

The research continues through early 2025:

Complete data collection — 60-75 projects across all categories
Implement full cohomology calculation — using GUDHI for proper simplicial homology
Fork prediction study — test the H² hypothesis on historical data
Write the thesis — formalize the categorical-cybernetic framework

Open Source, Open Research

The entire project is open source:

Repository: github.com/ibrahimcesar/categories-of-the-commons

Data collection scripts
Entropy calculation modules
Jupyter notebooks for analysis
The complete sheaf-theoretic framework (in theory/sheaf-cohomology-framework.md)

Why This Matters

Beyond academic interest, this research could have practical applications:

Early warning systems for governance crises
Diagnostic tools for foundation-backed projects
Better understanding of when to apply Ostrom’s principles vs. other approaches
Mathematical foundation for studying organizational health

If you maintain an open source project, especially a Stadium-type one, I’d love to include it in the study. If you’re interested in category theory applied to real-world systems, the theoretical framework might interest you.

This research is part of my MBA in Strategic Management at University of São Paulo (USP). It builds on my previous work on category theory for developers and my experience maintaining react-lite-youtube-embed.