Immutability is everything!
summary
Argues that making state immutable removes shared-memory headaches so concurrency becomes simple and reliable, demonstrated with Erlang/Elixir-style no shared memory processes that use message passing and copy-on-send. Claims this lets systems scale from Raspberry Pi to supercomputers (Sunway TaihuLight at ~ten million cores and ~93,000 trillion FLOPS) by trading latency for parallelism, and that distribution, replication and CRDT/eventual-consistency patterns handle availability and failure.
AI-written summary · about
chapters (11)
- 0:00The case for functional
- 2:01Concurrency and immutability
- 5:50Understanding immutability
- 8:29Mutable state challenges
- 11:51The power of no shared memory
- 14:44Distribution and latency
- 18:09The multi-core era
- 23:11Amdahl's law and scale
- 26:41Scalability and reliability
- 30:32Handling distributed failure
- 34:10Future of programming
related talks
01h 01mThe Power and Practicalities of Immutability
42mPure Functions and Immutable Objects
39mConcurrent Languages are Functional
39mConcurrency Oriented Programming In A Modern World
19mImmutability Matters
34mImmutability: Putting The Dream Machine To Work
43mConcurrency Abstractions for Application Security
52mImmaculate Code (pure code)