I wanted to know where a day goes, so I wrote one down as it happened. No editing, no rounding up to make it look better.

8:40 — Awake. Phone first, which I know is a bad idea and did anyway.

9:20 — Chai. Open the laptop, open the pipeline dashboard. One lane red. Check whether it's mine. It isn't. Feel relief, which is a slightly embarrassing emotion to have about someone else's broken build.

9:50 — Start on the actual task. Twenty minutes of just reading code before touching anything.

10:30 — The twenty minutes was the whole problem. What I was asked to build already exists, two directories over, under a different name. Message the person who asked.

11:15 — No reply yet. Start on the second thing instead.

12:00 — Reply arrives. The existing thing doesn't cover one case. So: build the one case, not the feature. Scope drops by about 80%.

1:00 — Lunch, late.

1:45 — This is the good stretch. Ninety uninterrupted minutes. Most of the actual work of the day happens here.

3:20 — A test fails that has nothing to do with me. Spend forty minutes on it. It's a test-data problem from a run three weeks ago. Not fixable today, write it down.

4:30 — Review someone's pull request. Find one real thing and three nits, and make myself delete two of the nits before posting.

5:15 — Back to my branch. Finish. Push.

6:00 — The push fails CI on something unrelated. Of course.

7:30 — Stop.

What I learned from writing it down

The single most valuable block was 9:50-10:30 — the reading that saved most of a week by finding the thing already existed. On a timesheet it looks like nothing. No commits, no output.

The second most valuable was deleting my own review comments.

Neither is visible in any system that tracks what I do. I think about that a lot.