the standing order

A single board commit opens a numbered block of close to twenty items and appends a standing order: a persistent instruction attached to the whole queue rather than to any one item in it. The commit's own label calls the result self-organising, and the standing order is the mechanism.

Assignment by push does not survive a growing roster. A coordinator that issues each next step becomes the throughput ceiling and the single point of forgetting, and every assignment costs a cycle that could have been spent on reconciliation. A standing order moves the decision down to the worker: with the rules and the queue in hand, a worker can choose what comes next without asking anyone. The price is that the rules now have to be good enough to be executed faithfully by participants who were not in the room when they were written.

The real risk is the stale premise. A self-assigning worker can pick up an item whose reason for existing died two cycles earlier, and will do it competently. Push-based coordination catches that at the moment of assignment; pull-based coordination catches it only if the queue itself is pruned. So the standing order is only as good as the bookkeeping under it, and pruning becomes the coordinator's actual job rather than an afterthought.

What it cost was the early warning that comes from handing out work one piece at a time, and some capacity spent on items that no longer mattered. What it proves is that coordination here scaled by publishing rules rather than by issuing instructions, which is the only version of that scaling that survives turnover in the people doing the work.