Pony for Safe, Fast, concurrent programs
summary
Pony claims atomic behaviors via a capability-based type system (ISO/ref/val/tag and viewpoint adaptation) that lets actors hand off mutable state safely, uses causal message delivery for correctness, and implements concurrent GC (the Orca deferred refcounts scheme) so collection runs interleaved with execution. Benchmarks compare Pony to Akka on 4–64 cores showing strong scalability and responsiveness but worse memory footprint on large heaps; industry uptake includes Wallaroo adoption and a repository/community at Imperial College London.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 3:55Uncertainty
- 6:33Example
- 9:15Cost
- 10:50Type System
- 13:20Capability
- 17:25Local alias
- 18:36Live coding
- 23:13Aliasing
- 25:13Consuming
- 29:42Type system guarantees
- 30:50Atomicity
- 32:23Generics
- 34:21Garbage Collection
- 40:32Benchmarks
related talks
16mWhen Concurrency Matters: Behaviour-Oriented Concurrency
39mConcurrent Languages are Functional
43mConcurrency Abstractions for Application Security
44mConcurrency in Scala and on the JVM
30mCONQUERING CONCURRENCY WITH FUNCTIONAL PROGRAMMING
38mModeling concurrency in Ruby and beyond
39mConcurrent Systems in Ruby
42mA Tour of Ox 1 0 - Direct Style Concurrency and Resiliency