The event loop finally clicked when I stopped calling it a loop

Tue, 18 Aug 2026 00:00:00 GMT

The event loop is often explained as a small circle that continuously checks a queue. That picture is useful, but it hides the part that matters most: JavaScript runs jobs to completion, and the host decides when another job may begin.

Once I started thinking in terms of turns instead of a literal spinning loop, ordering became much easier to predict.

A turn, drawn plainly

A tiny experiment

console.log('first');
queueMicrotask(() => console.log('microtask'));
setTimeout(() => console.log('task'), 0);
console.log('last');

The current script finishes first. Microtasks drain next. Only then can the timer task start a new turn.

The rule I keep nearby

That gives us a compact rule worth remembering:

Finish the current job.
Drain the microtask queue.
Let the host choose the next task.

The rule is smaller than most diagrams, but it explains more of the behavior I encounter in real code.

Debugging got faster when I started keeping an evidence ledger

Sun, 09 Aug 2026 00:00:00 GMT

The most expensive part of debugging is often not finding evidence. It is forgetting which conclusion came from which observation.

I now keep a tiny ledger with three columns:

Observation	Hypothesis	Next experiment
The retry happens after 30 seconds	A client timeout is firing	Log timestamps at both boundaries
The server finishes in 12 seconds	The response is delayed elsewhere	Compare proxy and browser timing

The format prevents a plausible story from quietly turning into a fact. Each experiment must either strengthen a hypothesis, weaken it, or reveal a better question.

This is slower for the first five minutes and dramatically faster for the next hour.

Intuitively — English

The event loop finally clicked when I stopped calling it a loop

A turn, drawn plainly

A tiny experiment

The rule I keep nearby

Debugging got faster when I started keeping an evidence ledger