Elm-CRDT is a Pure-Elm library implementing Conflict-Free Replicated Data Types, enabling concurrent editing with automatic convergence regardless of merge order or decentralization. It supports multiplayer collaboration, offline capability, complete edit history, and features like delta sync, collaborative text via Fugoe, and rich text via Peritext.
Respawn is a Rust-based content-addressed filesystem snapshot tool that enables versioning, undoing, and reverting directory states with cryptographic integrity guarantees. It supports atomic per-file reverts, tamper-evident auditing, and peer-to-peer replication over LAN without a central server, designed for safely rolling back changes made by autonomous processes.
A curated collection of seminal research papers in distributed systems, spanning foundational work on event ordering, Byzantine fault tolerance, consensus algorithms, and replication methods from 1978 to 2014. The list includes influential contributions from researchers like Leslie Lamport and covers critical problems that shaped modern distributed computing.
Walshadow is an experimental tool that replicates PostgreSQL rows into ClickHouse using physical WAL, supporting initial load, continuous changes, and schema evolution. It requires PostgreSQL 16+ and is currently in development preview with interfaces subject to change.
PostgreSQL 19 introduces WAIT FOR, a new SQL command that blocks a session until WAL reaches a specified position, enabling read-your-writes consistency on asynchronous replicas without the performance cost of synchronous replication. The command solves the stale read problem by allowing applications to wait for specific WAL positions before reading from standby databases.
GitHub is experiencing an active service outage affecting multiple services including API requests, issues, pull requests, Actions, and Pages. The incident is caused by increased database replication delays on the collaboration system, leading to higher error rates in authorization endpoints and cascading failures across the platform.
A Kafka cluster experienced severe performance degradation when eight single-member consumer groups, each fetching 4,742 partitions, caused lock contention in the broker's DelayedFetch mechanism. The issue emerged after upgrading to Kafka 2.7+, which changed lock acquisition to blocking mode, starving request handlers and blocking network threads while broker-measured latency metrics remained artificially low because time spent in unread sockets goes unmeasured.
Jetpack is an add-on framework for leader-based consensus protocols that enables 1-RTT fast-path commits by running parallel execution paths, reducing latency in WAN deployments. However, the dual-log architecture introduces redundant network traffic and CPU overhead, and the fast commit optimization only benefits write-heavy workloads with nil-externality; interactive reads still require the full 2-RTT latency. Evaluation across six consensus systems on AWS datacenters demonstrates gains for write-heavy pipelines using YCSB workloads.
Project Callisto, a nine-year research program (2015–2024) involving over 100 scientists, conducted extensive experiments to verify cold fusion claims from 1989. After systematic testing across electrochemical and ultrapressure loading conditions, the program found no evidence of anomalous heat effects and concluded that prior reports of excess heat have prosaic, non-nuclear explanations.
An engineer shares lessons from operating petabyte-scale ClickHouse clusters for nearly 6 years at Tinybird, covering architecture decisions including replica and shard strategies, HTTP-based load balancing, and cost considerations. The piece emphasizes that cluster setup is straightforward but operational challenges arise in scaling, storage management, and avoiding costly replication overhead.
Kafka Simulator releases v1.5 and v1.6 with 19 new scenarios, completing the cluster topology ladder with 3-DC stretched and 2.5-DC stretched cluster shapes. The updates add storage/lifecycle management and ops/controller capabilities, enabling users to test multi-region architectures including synchronous replication across datacenters and witness-based quorum designs.
Neki is PlanetScale's sharded Postgres solution that scales beyond single-machine limits to handle hundreds of millions of queries per second and petabytes of data. It maintains 100% real Postgres in each shard while adding a router, sidecars, and control plane for horizontal scaling, with zero-downtime resharding and online migrations.
A scientist critiques the growing tendency among researchers to excuse questionable practices by blaming perverse incentives in academic publishing and career advancement. Rather than accepting this as inevitable, the author argues that scientists have agency and could make ethical choices despite systemic pressures, but increasingly use 'the incentives' as a convenient rationalization for avoiding difficult decisions.
Multigres v0.1 alpha is an open-source operating system for PostgreSQL that provides horizontal scaling, high availability, and operational management through sharding, connection pooling, automatic failover, and backup orchestration. The release includes a Kubernetes operator, advanced connection pooling with intelligent traffic routing, and a consensus-based HA system that handles split-brain scenarios while preserving committed transactions.
Peer review became the standard for scientific publishing after World War II when government funding surged, but evidence suggests this massive, costly intervention has failed to improve research productivity or quality over the past six decades. Scientists now spend 15,000 years annually reviewing papers, yet many peer-reviewed findings don't replicate and new ideas struggle to displace established ones.
A 2002 paper on procrastination and deadlines coauthored by Duke behavioral scientist Dan Ariely has been retracted after Data Colada analyses found the underlying data had been tampered with. The retraction, the second for Ariely, followed a failed replication attempt and comes 24 years after the paper's original publication in Psychological Science, where it became highly influential with nearly 1,000 citations.