A New Programming Model for Full-Stack Distributed Systems
summary
A new framework called reboot.dev implements a concrete programming model (TypeScript and Python) that treats data as durable, addressable objects described in protobuf IDL and exposes typed functions as readers, writers, transactions, workflows. It generates React hooks like useAccountBalances, enforces semantics (double-execute strict mode in dev), makes idempotent calls by default, and offers atomic transactions across partitioned instances; the implementation exists now and is downloadable so the demo claims can be tried.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 0:26Background and Motivation
- 2:27The Reboot Framework
- 6:06Reboot Data Model
- 8:55Reboot Functions
- 10:39Effect Management
- 13:34Enforcing Semantics
- 15:01Imperative and Functional
- 17:06First-Class Reactivity
- 21:22Handling Unreliable Networks
- 24:18Item Potency in Practice
- 25:59Data Inconsistencies
- 30:03Database Assimilation
- 32:28First-Class Transactions
- 35:14External Effect Challenges
- 38:15Reboot Workflows
- 46:15Summary and Future Vision
- 47:44Q&A Session
related talks
32mReactive Data on the Full Stack
46mDevelopment Without Front-Tiers
45mThe simplest way to build resilient applications
25mEnabling Distributed Transactions Across Microservices Using gRPC
01h 03mThe Power of the Paradigm
58mFrom Zero to Live with Functional JavaScript Applications (Part 2)
40mIt's Time to Rebuild DevOps
30mDX is not magic, and neither should your backend be!