Aspect-oriented Architecture
summary
Argues that architecture becomes code by running concrete architectural fitness functions (examples: ArcUnit/ArcUnit.net, pytest-arc, TSARC) to enforce rules such as no cyclic dependencies, layered-access constraints, and GitHub actions that warn on feature-branch age. Describes an architecture definition language used with generative AI to produce platform-specific tests, a Ruby example that imports logs to verify an orchestrator is always called, data checks that hash ticket keys until P=0 to ensure consistency, and the emerging agentic mesh pattern. Pull requests grow ~10% from genAI.
AI-written summary · about
chapters (17)
- 0:00Aspect-oriented architecture intro
- 1:06AOP origins and flaws
- 3:38Architectural fitness functions
- 7:29Preventing cyclic dependencies
- 10:59Compile-time fitness limits
- 13:31Fitness function examples
- 15:37Architecture Definition Language
- 17:45Generating platform-specific fitness
- 21:41Enforcing layered constraints
- 24:53Data consistency fitness
- 27:46Monorepo governance
- 31:59Feature branch longevity
- 33:07Hexagonal architecture revisit
- 36:20Data mesh patterns
- 37:42Agentic mesh trends
- 39:51Fitness function summary
- 40:33Q&A
related talks
25mThe Intersection of Fitness Function driven Architecture and Agentic AI
48mDown the Ivory Tower towards Agile Architecture
58m4 Key Trends Towards Better Software Architecture
58mThe Intersection of Architecture and Implementation
11mClean Architecture and AI
44mAI Architecture Thinking: From Idea to Impact
22mArchitecture & ArchUnit
49mBottom-up Architecture: Bridging the Architecture-Code Gap