Web Workers

What you'll learn

• When the main thread is the wrong place — and how to tell
• Dedicated workers, and the postMessage protocol that joins them
• Transferable objects, and why they're not optional for big data
• Shared workers and the cross-tab use cases they enable
• Service workers vs web workers — they're different
• Worker pools, RPC patterns, and structured cloning costs
• Debugging workers in DevTools

1. 01
   Why workers exist

   The event loop, jank, and the cases nothing else fixes.
   13 min
2. 02
   Dedicated workers

   Spawning, messaging, terminating — the lifecycle in 60 seconds.
   14 min
3. 03
   The postMessage protocol

   What's structured-cloned, what isn't, and the cost.
   15 min
4. 04
   Transferable objects

   Move, don't copy — ArrayBuffer, MessagePort, OffscreenCanvas.
   14 min
5. 05
   Shared workers

   One thread, many tabs — the cross-tab broker pattern.
   13 min
6. 06
   Service workers vs web workers

   Same word, different jobs. Don't confuse them.
   11 min
7. 07
   SharedArrayBuffer and Atomics

   Real shared memory, COOP/COEP headers, and the locks you need.
   14 min
8. 08
   Worker pools and RPC

   From one-shot tasks to a pool — and a tiny RPC layer.
   14 min
9. 09
   When workers aren't the answer

   Cost of message passing, when to scheduler.yield instead.
   11 min
10. 10
    Worker debugging

    Sources panel, separate consoles, and finding errors that vanish.
    11 min