Convergent/Divergent

on YouTube ↗33m

" summary

If an Ember.js app pushes state to clients it already behaves like a distributed system, and the concrete fix is to embed causality into data structures: explain convergent and commutative replicated data types, contrast state‑based versus op‑based CRDTs, and show the observed-remove set and lattice merges where union/max/OR give associativity commutativity idempotence. The claim is that mathematically proven merge functions let replicas converge without global consensus; implementations include counters, sets, registers and composable maps and they work in practice.

AI-written summary · about

" chapters (7)

1. 0:00Introduction
2. 4:57Building distributed systems
3. 9:09Consistency models
4. 12:20Consensus challenges
5. 17:35Conflict-free replicated data types explained
6. 27:20Why CRDTs matter
7. 28:44Conclusion

" related talks

• 36mCRDTs IllustratedArnout Engelen · Strange Loop 2015
• 45mBuild better apps using CRDTsEric Meier · JavaScript fwdays’23
• 32mCRDT and other new ideas for client-server communicationAndrey Sitnik · EmpireJS 2018
• 40mDistributed, Eventually Consistent ComputationsChristopher S. Meiklejohn · Strange Loop 2015
• 27mOn the Development of Reactive Systems with Ember.jsClemens Müller, Michael Klein · EmberFest 2021
• 52mState management beyond the librariesMichel Weststrate · HolyJS 2018 Moscow
• 29mCRDT: Conflict-free Data Types for Collaborative Editing and Offline-FirstAndrey Sitnik · NEJS Conf 2018
• 36mKnowing What You See - Let's Talk About Front-end ConsistencyYifan Wu · Strange Loop 2016