The Long Field

I. The argument

The building at Weena 200 has forty-one floors, three thousand occupants and, according to its facilities team, one recurring emergency. It is not fire, or flood, or the lifts. It is the eleventh floor being too cold while the twelfth is too warm, on the same afternoon, from the same air.

I have now sat in enough of these meetings to recognise the shape of them. Someone from facilities explains that the system is operating within specification. Someone from the eleventh floor produces a photograph of a colleague in a coat. Everyone agrees that something should be looked at. Nothing is looked at, because nothing is broken.

That last part is the interesting part. The building is working exactly as designed, and the design is the problem — not because the engineers were careless, but because the thing they were asked to optimise is not the thing the occupants experience.

II. The standard

The specification in question is a comfort standard, and comfort standards are among the strangest documents in engineering. They are attempts to write down, in degrees and percentages, a state of not-noticing.

The dominant model was built in the 1970s from experiments in a climate chamber in Denmark. Volunteers in standardised clothing sat in a controlled room and reported how they felt on a seven-point scale. From thousands of these votes came a curve that predicts what proportion of people will be dissatisfied at any combination of temperature, humidity, air speed and clothing.

It is a genuinely brilliant piece of work, and it contains a quiet admission that has shaped every office building since: the curve never reaches zero. At its very best, in the most precisely controlled conditions the model can describe, five per cent of people are still uncomfortable. Comfort, formally defined, is a state in which one person in twenty is unhappy and the building is considered to be working.

Comfort, formally defined, is a state in which one person in twenty is unhappy and the building is working.

III. The override

What happens next is the part that never appears in the specification. People adapt the building by hand.

In the towers I have surveyed, roughly one desk in nine has a personal heater, a fan, or both. Windows that are designed not to open are opened. Sensors are covered with sticky notes, blocked with cardboard, or — in one memorable case in Utrecht — carefully warmed with a desk lamp positioned eleven centimetres away, which persuaded an entire zone that it was two degrees warmer than it was and turned the heating off for a floor of two hundred people.

Facilities teams describe these acts as misuse. I would describe them as data. Every space heater under a desk is a measurement: someone has decided that the discomfort of the building exceeds the social cost of visibly disagreeing with it, and has spent their own money to say so.

The building, meanwhile, reports that it is operating within specification, and it is telling the truth.

IV. What follows

There is a version of this essay that ends with a demand for better sensors, and I do not want to write it. More sensors would tell us with greater precision that the average is fine, which we already know.

The more interesting shift is happening in a handful of buildings that have stopped treating the override as a fault. They allow openable windows on a floor that also has mechanical cooling, and accept the small energy penalty. They give occupants a control that actually does something, rather than a placebo dial. They measure how often people intervene and treat a rising number as a design signal rather than a training problem.

None of this is technically difficult. What it requires is a change in what the building is understood to be for — from a machine that produces a specified climate to an instrument that a few thousand people play, badly, all day, and that is designed with that in mind.

The eleventh floor will still be cold. But someone will be allowed to do something about it, and the building will be listening when they do.