The file called final

Every office archive eventually contains a file called ‘final’. It is soon joined by ‘final_v2’, then by ‘final_v2_FINAL’ and at last, in a spirit of typographical despair, by ‘final_v2_FINAL_use_this’. None of these names is untrue at the moment it is typed. All of them are untrue a week later.

I do not think archives fail through laziness. They fail because a file is named when less is known about it than at any later point: the moment it is saved. Whoever types ‘final’ is expressing a hope, not describing the document. Folders go the same way. They are founded on one principle, by year or by subject, and grow according to another, until the folder called ‘Misc’ is the largest and the most consulted, which says a good deal about the candour of the rest.

The result is an archive that keeps everything and returns nothing. An organisation’s memory is not the sum of what it has saved. It is what it can find in five minutes, on the day it is needed, while the one person who knew where things were is on holiday.

An archive in which things can be found again asks, in my experience, for four habits. None of them is technological.

The first is a fixed address. Each kind of document has one place where it lives, and that place does not move because somebody’s mood has. Copies exist, but they know they are copies.

The second is an honest name. A name says what the file contains and which date it refers to, not what anyone hopes of it. ‘2026 accounts, draft for the March review’ is less stirring than ‘final’, but every word of it is true.

The third is a source for every claim. A figure in a presentation ought to be able to say where it came from. If the answer is ‘somebody’s email, July, I think’, the figure is not a fact. It is a hypothesis in a good suit.

The fourth is the most neglected: what has been superseded stays, but says so. Deleting old versions looks tidy; in practice it destroys the record of how a decision was reached, which is often the very thing one is looking for. It is enough for the superseded document to say so on its first line and to point to its successor. An archive that marks its old versions is an archive. One that does not is a treasure hunt without the treasure.

None of this needs new software. It needs a decision, taken once, about where things live and what they are called, and the discipline to keep to it when in a hurry, which is to say always.

As for the file called final, I have seen one survive three reorganisations. It is the only truly final document in the office. Nobody remembers what it is about.