Weeknotes for the week finishing Friday, 24th April 2026.

This week, I:

  • side-stepped a problematic dependency update in OPv1.
  • thought about overhang.
  • asked “What’s the point of code review?”

Overhang

Last week was the final week of our current cycle, during which we added dm+d forms/routes and ingredients to OPv2. Finishing cycles is hard. We had a lot of overhang, and we spent this week tidying up.

Spending a week tidying up isn’t necessarily a warning sign, but this time the overhang included a large, AI-assisted PR.1 It seemed as though if we deployed the PR ready for the show-and-tell, then we’d have to accept that we wouldn’t review it. But if we reviewed the PR, then we’d have to accept that we wouldn’t deploy it. Not deploying it would mean our colleagues couldn’t click along with us during the show-and-tell. It would also mean we’d have to rethink who would be telling and what they would be showing at short notice.

What did we do? We agreed to a post-deploy review. We would deploy the PR ready for the show-and-tell, stick to our original who/what plan, and review the PR later. At least, that’s what we agreed on Friday. On Monday, we still hadn’t deployed the PR! So, we made a new who/what plan at short notice. We talked about the work on the PR, but we showed it from a local development branch. And rather than being the focus of the show-and-tell, it was one new feature among three(ish) new features.

What happened next? We deployed the PR on Tuesday. We didn’t review it. We probably won’t. Hence my question: “What’s the point of code review?” I started making some notes, which I hope to write up.


  1. It touched roughly 2,000 lines of code.