Effective Reactive Architecture
summary
Shows how an Rx-based UI moves from tangled subjects to a single unidirectional stream: view events -> business results -> immutable view models -> view rendering. Implements optimistic updates with startWith, uses the scan operator to fold results into state, exposes presenter logic as an ObservableTransformer (events in, view models out), and preserves presenter state across rotations with replay + autoConnect(0). Recommends subjects only when lifecycles differ, keeps side effects isolated, and handles navigation as a doOnNext/ignoreElements side-effect.
AI-written summary · about
chapters (9)
- 0:00Introduction to reactive architecture
- 1:04Side effect isolation principle
- 6:20Refactoring messy RX code
- 13:21Presenter design patterns
- 20:22Handling configuration changes
- 25:21Unidirectional data flow
- 27:07Adding new features cleanly
- 30:57Navigation in reactive apps
- 33:04Summary and key takeaways
related talks
28mApplying Rx Best Practices to Your Architecture
35mModel View Intent, Embracing Reactive UI's
36mRxJS: A Better Way to Write Frontend Applications
51mManaging State with RxJava
33mStop Micromanaging State and Start Being Reactive
34mOperate Heavily Dynamic UIs with RxJS
40mReactive Programming: A Better Way To Write Frontend Applications
45mGoing Reactive - An Introduction to Reactive Programming