Main idea

Three statements were put forward in turn as proof that a product reference model was finished. "The card network is closed": every card cited exists. "I've been through all my sources": the code, the database, the API, the glossary, the interviews, the support tickets — nothing left to open. "Every area has been read in all four places": the core code, the database, the screens, the translations.

All three are true. All three prove the same thing, and it isn't the right one: I have finished what I started. None of them says nothing is left. "I've read my whole stack" says nothing about the library: until the object of the work is enumerated, "everything" has no denominator.

This isn't dishonesty, it's a matter of cost. A satisfiable criterion is already at hand: you just look at what you've done and note that it's done. A criterion that proves something requires manufacturing something extra — the list of what exists, a way of counting it, a number someone else could recompute. That is additional work, demanded at the precise moment when you believe you have finished.

Why it matters

The distinction serves well beyond a reference model: an audit, a migration, a security review, a technical-debt cleanup almost always end on "I've looked at everything". The question to ask is always the same: everything, meaning what? Do you have the list?

It is also what lets you refuse a proof of completion without calling into question the seriousness of whoever presents it: you aren't attacking the work, you are noting that the sentence bears on the work and not on its object.

Nuances and limits

A satisfiable criterion isn't useless: "the network is closed" remains a good cohesion check. It only misleads when you make it say what it doesn't measure.

And manufacturing the denominator isn't always possible: some objects of work can't be enumerated. In that case you have to accept that no criterion will prove anything, rather than settle for one by default.

Open questions

• Who, in an organisation, has the mandate to refuse a stopping criterion, given that the refusal extends a piece of work everyone believed was finished?