What is the difference between a controlled document and a working file?

Why the distinction is not pedantry

The machinery of a document register — numbering, revision control, status codes, controlled distribution — exists because somebody is going to build from the document, and everybody needs to know they built from the right version of it. That machinery is expensive to operate, and it earns its cost precisely because it is applied to the things that need it.

Put every file through it and you have not made the project more rigorous. You have made the register useless: four thousand entries, of which two hundred are things anybody will ever build from, and a revision-control discipline that now means nothing because it applies to a supplier brochure.

Where each one belongs

Two tiers, and the sorting is usually obvious once you ask the right question — will somebody act on this as though it were authoritative?

• Controlled: drawings, specifications, contracts, issued reports, approved method statements. Numbered, revisioned, statused, distributed on the record.
• Working: drafts, calculations and their workings, site photographs, meeting decks, supplier brochures, the spreadsheet somebody built to check somebody else’s numbers.
• The trap: leaving the working tier ungoverned. It does not stay uncontrolled — it goes to personal drives, email attachments and chat threads, where it is invisible to the project and leaves with the person who made it.

Crossing the boundary

Files mature. The draft becomes the issued report; the markup becomes the revision. At that point the file stops being a working file and becomes a controlled document, picking up numbering and revision control from then on, with its working history preserved.

That crossing should be a deliberate act, and a human one. A document becomes controlled at the moment somebody is prepared to be accountable for it — and a system that decides that on its own has quietly taken the decision away from the person who will have to answer for it.