Mesh
summary
A new experimental language called Mesh targets practical concurrency by combining serializable STM over persistent data structures, a tupled function model with type inference, and control-as-library combinators like a parallel-for combinator and spawn. Demos show a Flickr thumbnail downloader doing synchronous network calls in parallel without callbacks, a lightweight spinner, a billiards simulation with runtime-adjustable thread count, and a scalable Game of Life; most transactions are tiny assignments, retries handle the occasional dynamic multi-cell update, and the project is planned to be open sourced by the end of the summer.
AI-written summary · about
chapters (20)
- 0:00Start
- 3:18The Sweet Spot(s)
- 4:55Starting points
- 6:30Q: Why so little concurrent code?
- 8:08Reducing concurrency lock-in
- 9:47Mesh STM model-genealogy
- 13:35Mental model
- 14:53Just act natural
- 16:17Demos
- 18:52Concurrency overview
- 20:36Parallel download and save
- 23:44Parallel-for
- 25:17Curly braces == lambda
- 25:27pfor body factored to function binding (1)
- 26:17Where's the transaction?
- 26:43Synchronous Background Networking
- 28:44While
- 36:22Ball update and cell transfer
- 39:43Assign
- 41:34Variadic signatures
related talks
01h 06mTransient: Unrestricted Algebraic and Monadic Composability
41mProgramming Distributed Systems
26mBuilding Cache Coherent, Heterogeneous-ISA Processors Using P-Mesh
39mConcurrency Oriented Programming In A Modern World
19mWe Know I Know You Know; Choreographic Programming With Multicast and Multiply Located Values
38mMesomatic: The Cluster is a Library
20mEfficient, Portable, Census-Polymorphic Choreographic Programming
23mA Haskell Auto-Parallelizer for Distributed Computing