Beyond REST: Using Full-Stack Signals for Real-Time Reactive UIs
summary
Argues that real-time UIs are easiest when treated as shared UI state and demonstrates a working poll: front- end signals (React/Angular/Vue or Preact signals) map to server-side Java numberSignals, updates go over WebSockets as increment-by operations to avoid lost increments, and concurrent votes converge correctly. Backend is initially in-memory with locking, plans an append-only event log and distribution via Hazelcast or Kafka for offline catchup and clustering. The demo works and preserves ordering and correctness under concurrent updates.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 1:07Realtime update use cases
- 3:07The key trick
- 4:03The poll
- 4:59How to build your UI
- 6:33Have you already built realtime UIs
- 7:35Why not
- 8:55Three basic steps
- 9:54UI Design
- 11:03Identify Go
- 12:03Secret Admin UI
- 19:15Signal Theory
- 22:12Signal Operations
- 23:31Latency
- 26:51Synchronization
- 32:48Questions
- 36:53Summary
related talks
45mBeyond REST: Using Full-Stack Signals for Real-Time Reactive UIs
52mIt's Time To Talk About Signals
40mReactive Programming: A Better Way To Write Frontend Applications
34mBeyond Signals
30mUnderstanding the realtime web with websockets
46mEvent-Driven UX in the Real World with Angular and Socket.io
30mAre Signals worth the hype?
32mVue Real-Time Apps