How the log was kept

The columns

Collisions took day, approximate time, room, the arriving activity, the departing one, and which person relocated.

Gains took day, time, room, what occurred, and which person.

Entries went in immediately where that was practical and within the hour otherwise — roughly twenty minutes daily, counting an evening pass to tidy them.

The two tests

Three conditions had to hold for a collision: two unlike uses, a room already in service, and one party relocating. Each is a matter of observable fact rather than interpretation.

A gain required that the episode occurred and that an office day would have prevented it outright — not merely complicated it.

Six gaps

Avoidance leaves no mark. Choosing not to enter a room because a call is audible through the door produces no entry. One cross-check on day nine turned up three relocations that the person making them had not registered. Treat forty-seven as a lower bound.

Quality is invisible. Half an hour at a folding table registers identically to half an hour at a proper desk.

Hours weigh the same. Six entries fall after six in the evening, and the household rates those as the costliest of the lot.

Sorting came later. Every category on this site — the four kinds of use, six kinds of gain, four causes of conflict — was derived from the notes once the fortnight had finished.

Days one and two are sparse. An initial attempt to record all room use, rather than exceptions only, collapsed on day three.

Energy went uncounted. No meter was read. The shift of heating cost from employer to occupant, which dominates public discussion of this subject, is missing from a log otherwise rich in room-level detail.

What a second attempt should do differently

Fix the weights before counting rather than after. Record how often each room changes hands. Read the meter twice. Find a second household.

In short

One household, one fortnight, written down by people with no case to make, and the sole assertion attached to it is that they wrote it down.