Architecture Beyond Clean & Hexagonal
summary
Argues that common patterns (MVC, layered, hexagonal/clean) miss a critical separation: pull logic into leaf operations and separate integration as its own concern. Uses a concrete CSV→histogram example with reader/parser/histogram classes, JUnit + mock readers for tests, and shows how IOSP lets targeted unit tests without dependency inversion. Proposes the Yoda architecture and Sleepy Hollow architecture for orchestrated processors and heads, and recommends event sourcing and vertical slices to scale — claiming simpler code, easier tests, and fewer injected abstractions.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 3:58Clean vs hexagonal rise
- 8:33Logic to functions
- 18:59Separation of concerns
- 30:29MVC revolution
- 37:40DIP limitations
- 46:20Hexagonal and clean explained
- 53:04Integration Operation Segregation Principle (IOSP)
- 1:00:35Yoda architecture
- 1:11:34Functional core pattern
- 1:14:03Sleepy Hollow for testability
- 1:20:03Vertical slice architectures
- 1:23:33Event sourcing models
- 1:26:29Q&A
related talks
51mA Contrarian View of Software Architecture
52mEvolving a Pragmatic, Clean Architecture A Craftsman's Guide
42mStratified design and functional architecture
41mUsing Boundary-Driven Development to beat code complexity
19mHexagonal Architecture in Angular
40mSOLID Architecture in Slices not Layers
25mTestable Architecture: Keep 'em Separated
50mTowards an Evolutionary Architecture