Git Archaeology: Fighting Code Entropy with Engineering Impact Scores

Entropy: The Universe Always Tends Toward Disorder

Machu introduces the Engineering Impact Score (EIS) framework to quantify the Second Law of Thermodynamics within codebases. The system identifies that every new feature addition inevitably increases complexity and disorder, requiring active energy to maintain structure.

Why This Matters

While ideal models suggest “don’t touch working code,” the technical reality is that code rots as its surrounding environment evolves. Maintaining order requires continuous energy expenditure; without it, even beautiful designs succumb to urgent shortcuts and ignored conventions, leading to structural collapse as entropy accumulates over time.

Key Insights

• Production as Entropy: Writing new code increases complexity and dependencies, driving the system toward chaos unless balanced by quality-focused activities.
• Quality as Resistance: High Quality scores in EIS, derived from tests and reviews, serve as the primary mechanism for resisting the natural decay of code order.
• Survival Metric: A Survival score of 100 serves as a benchmark, proving a specific segment of code has successfully resisted erosion for at least six months.
• Design as a Barrier: Architectural boundaries act as physical walls that limit the blast radius of changes and block the propagation of entropy across modules.
• Breadth for Early Detection: Engineers with high Breadth scores monitor a wider range of the codebase, allowing them to sense signs of rot before they manifest as failures.

Working Examples

Installation command for the Engineering Impact Score (EIS) CLI tool.

brew tap machuz/tap && brew install eis

Practical Applications

• Use case: Architects with high Design scores build entropy barriers to isolate legacy rot. Pitfall: Neglecting refactoring leads to dark matter work that doesn’t show in logs but causes eventual system collapse.
• Use case: Monitoring EIS timeline trends to detect declining Quality scores before survival rates drop. Pitfall: Treating working code as immutable, which allows its relative order to degrade as the surrounding environment changes.

References:

• https://dev.to/machuz/git-archaeology-11-entropy-the-universe-always-tends-toward-disorder-ak9