An Introduction to Residuality Theory
summary
Residuality theory models software as sequences of leftover systems — residues — and claims architects should train systems to survive unknown events rather than predict them. It uses random simulation (Monte‑carlo style darts) plus Calman random Boolean networks (n=100,000 → ~317 attractors observed) to find attractors, then an incidence‑matrix contagion analysis to expose invisible coupling. Example: an EV‑charger system adds ALPR, redundancy and billing changes and survives messy Xs like “icing” and vandalism. A measurable residue index (y−x / #stressors) proves improvement.
AI-written summary · about
chapters (17)
- 0:00Software architecture challenges
- 1:28Time change in architecture
- 3:55Missing unit of architecture
- 7:01Ordered vs disordered systems
- 10:30Complexity science in software
- 13:18Random simulation in systems
- 14:21Kauffman networks and attractors
- 21:53Two-step algorithm in engineering
- 23:40Residuality theory basics
- 26:31Complex business environments
- 29:18Residues as architecture unit
- 33:56Stress testing with examples
- 39:59Criticality over correctness
- 43:30Contagion analysis
- 48:28Proving architectural improvements
- 50:55Residuality vs traditional methods
- 53:03Summary and learning
related talks
47mResidues: Time, Change & Uncertainty in Software Architecture
39mModel Drift and Software Attractors
26mBreaking Systems for Resilience
52mThe Philosophy of Architecture
41mHow Programs Learn, and What Happens After They're Built
44mFractal Geometry of Software Design
44mBuilding resilient frontend architecture
57mResponsibility Driven Design Revisited