Functional Async on the JVM

on YouTube ↗53m

" summary

Claims a practical abstraction for async/parallel functional Scala on the JVM built as a monadic Task with now/delay/async that respects JVM constraints (no TCO, expensive threads ≈10,000 limit, costly object churn). Implements a free‑style algebra and an interpreter that encodes trampoline encoded TCO and conserves threads, uses unsafe run with a latch to block only at the world edge. Works for streaming (S2/FS2) and async IO but exposes unsafe edge cases (double callbacks, nontermination, type‑soundness traps).

AI-written summary · about

" chapters (12)

1. 0:00Introduction
2. 1:03JVM and Scala async context
3. 6:02Tail call elimination limits
4. 8:10Threading and GC limits
5. 13:10Functional async abstraction design
6. 26:09Monadic task structure
7. 31:31Free monad task implementation
8. 38:32Unsafe task execution
9. 44:51Practical async benefits
10. 49:07Async drawbacks and safety
11. 51:10Task implementation pitfalls
12. 52:13Q&A

" related talks

• 01h 12mAsynchronous and Concurrent Programming in Scala Part 3John DeGoes · LambdaConf 2019
• 01h 53mAsynchronous and Concurrent Programming in Scala Part 2John DeGoes · LambdaConf 2019
• 58mBlazing Fast, Pure Effects without MonadsJohn A De Goes · LambdaConf 2018
• 01h 28mAsynchronous and Concurrent Programming in Scala Part 1John DeGoes · LambdaConf 2019
• 48m10 tips to write fast stack-safe functional code in Scala: the automatic-batch…Christophe Calvès · ScalaIO 2019
• 58mMonads in ChainsIvan Čukić · Meeting C++ 2014
• 47mWhirlwind Tour of the Scala Functional Programming EcosystemMark Moore · LambdaConf 2019
• 45mZIO: Next-Generation Effects in ScalaJohn A De Goes · ScalaIO 2018