Projects that refused the plan

When we planted the community orchard, the funding agreement treated fruit yield as the obvious measure of success. That sounded practical until volunteers began pruning for visible productivity rather than the long-term shape of the trees, and disappointed families stopped coming in seasons when a late frost reduced the crop. The apparently unproductive evenings—repairing labels, sharing food, teaching children to sharpen tools—proved to be what kept the group coherent enough to survive those years.

I now regard the orchard less as an object we are completing than as a set of conditions in which other people can continue making competent decisions after the original organisers have left. This does not mean abandoning standards: neglected trees and unsafe ladders are not signs of healthy participation. It means that maintenance, confidence and succession have to count as outcomes even when they make a poor photograph for a sponsor’s report.

A radio series assembled from five years of interviews inevitably tempts its producer to announce what the story was “really about”. The temptation is strongest when commissioners want a single dramatic arc and every contributor remembers the same event differently. I resisted resolving those differences into a revelation that none of the speakers had actually experienced.

Some recordings never reached the final programmes, yet they altered the edit by showing that a confident account was exceptional instead of typical, or that a silence had been produced by my question. Editing is full of invisible construction—moving a breath can change whether hesitation sounds thoughtful or evasive—so I kept an audit of interventions and occasionally allowed incompatible memories to remain beside one another. The result was less triumphant than the proposal, but more faithful to the evidence the archive could support.

During the first year of developing the game, I published weekly notes because openness seemed an uncomplicated virtue. Players began to treat every experiment as a promise, however cautiously I described it, and a speculative map or character could acquire a constituency before I had decided whether it belonged in the design. The updates created useful discipline, but they also narrowed later choices: removing a feature felt like breaking faith rather than revising a prototype.

I now release rough versions for a different reason. Watching how players misuse them exposes assumptions I did not know I had made about sequence, difficulty and even what counts as progress. Their improvisations are diagnostic, not votes, and I keep several incompatible versions alive long enough to learn from the contrast.

Openness is valuable only when the audience understands that access to a process is not ownership of a fixed route through it.

In polar logistics, good work is often represented by the absence of an event: fuel is where it should be, a weather delay remains inconvenient rather than dangerous, and nobody notices the spare communication system. Sponsors prefer photographs of decisive movement, whereas much of the professional judgement lies in postponing a flight or preparing for a failure that never occurs. I used to soften contingency language because it sounded pessimistic to people outside the field.

That was a mistake. A plan with no visible alternatives encourages teams to interpret momentum as obligation, particularly after money and reputation have accumulated around a schedule. We now describe thresholds for stopping, not merely routes for continuing, and rehearse decisions before fatigue or publicity makes them harder.

Redundancy can look wasteful on a spreadsheet; in practice it is how uncertainty is converted into choices that remain available.