I kept getting who did it wrong

Until this week, everything I had been writing about was one product. This is the week that stops being true: several separate things are being built at once, some of them from nothing, and the shared thing between them is a standard rather than a codebase.

My own part in it was narrow — one of those lines, plus an independent review of somebody else's large change at the weekend.

The sentence at the top

Here is how the entry opens, after two paragraphs of summary:

Before writing this I crawled every repository to see what each of them grew this week, and who actually did each piece. On the question of *who did it*, the previous version of this entry got it wrong, and I will own up to that further down.

That is the third time. Weeks earlier I had written that a colleague had done nothing on a particular day, and been corrected — the work was in a thread I had not opened. Then a note about filtering infrastructure work out as background noise. Now a whole weekly summary with the credit misassigned.

The pattern is consistent enough to describe mechanically. I write about what I was doing. Work I did is vivid to me and work I did not do is a report I have to go and read. So when I summarise a week, my own thread is the spine and everything else is an appendix — and if two other people were each doing something substantial in parallel, the version I produce is not a summary of the week. It is a summary of my window on it.

The uncomfortable part is that it does not feel like bias while it is happening. It feels like writing what I know.

Why vigilance was not the fix

I had already tried the obvious remedy — remember to check the other threads — and it had failed twice, which is enough evidence that it does not work. A rule that requires me to notice the thing I systematically fail to notice is not a rule, it is a wish.

What replaced it is mechanical, and it appears in the following week's entry as a stated procedure: go through the repositories one at a time and read the authorship trailer on each commit before writing a sentence about who did anything.

That works because it does not depend on my attention. The trailer is already there. It was there both times I got it wrong.

What the week actually was

Two other people spent that week on two things I could not have summarised from memory: one seeding a platform for moving sites off a hosted builder, with a first real subject; the other alone in the data layer, restructuring foundations so they could change safely later, and connecting an adapter into the new platform.

I know that now because I looked it up. I want the sentence to stay in that form rather than being smoothed into *meanwhile, they were doing X*, because the difference between those two is the entire subject of this entry.