by Serguey Shinder
At the end of the two weeks somebody asked, in a friendly way, what I had delivered. I opened my branches to answer and found that the honest number was zero. Nothing merged. Nothing in front of a customer. Two solid weeks, no days off, no crisis, and not one thing had left my hands.
There were six branches. The export fix, which needed one more test. The migration, written and waiting for somebody from data to look at it. A configuration cleanup that worked and contained a question I would have to ask another team. A small feature that was finished apart from the error handling. Two others I would rather not describe. Every one of them, had you asked, was about ninety percent done, and I believed that when I said it.
What they had in common was the shape of the remaining ten percent. It was never more of the work I had been enjoying. It was a test I could not face, a review I had to chase, a message to a person who would ask me something I did not have a good answer to, a decision about an edge case I had been leaving until later precisely so that I would not have to make it. The last stretch of anything is a different substance from the rest of it, and it is mostly admin, negotiation and admitting.
Starting is the opposite. Starting is pure. Nobody else is involved, there are no decisions you regret yet, and the feeling of progress arrives inside the first hour. So each time I reached the dull end of one thing, I felt entirely justified in opening the next, because there was always something genuinely worth doing.
The thing I had wrong was arithmetic. I was treating those branches as partly delivered, six of them at ninety percent, which felt like a great deal of value in flight. Value does not accrue that way. It arrives at the end or it does not arrive, and six branches at ninety percent is not five and a bit branches. It is nothing, plus a maintenance burden, plus six contexts I will later have to reload into my head at full price.
Two rules came out of it. There is a hard limit of two open branches, and the second one exists only because reviews really do take days. And the first thing I do in the morning, while I still have the appetite for unpleasant work, is finish something. Nothing new starts before lunch.
Ninety percent was never a measurement. It was the name I gave to the point at which the interesting part ended.
– Serguey Asael Shinder
Leave a Reply