An Aging Counter Is a Measurement, Not a Mechanism

62 open questions. 44 of them overdue. 20 sitting between 70 and 73 days old.

That is not a broken tracker. That is a tracker doing exactly what it was built to do.

Why the Register Exists

Every project has the question that gets asked in refinement, half-answered in a corridor and never written down anywhere. Three weeks later it decides a test result, and by then nobody remembers there was a question at all.

So the questions got a register. Auto-generated from the analysis files, one ID per question, an age in days on every row. Nothing can quietly fall out of it.

That was the design. The design works. Every row in that regeneration is accurate.

The Delta Line

The weekly health check prints a delta against the previous run. This week it read: all five next steps from last week are still open, ages advanced by 7.

Read it twice. The tool's own summary of a week of work is that a week passed.

Nothing moved. The register noticed, recorded the fact precisely, and had no further opinion about it.

Eleven Weeks, Three Questions, Three Suites

Three baseline questions gate the ACTIVATE, AMEND and CANCEL suites. Until somebody answers them, those suites cannot be written, because the expected results depend on the answers.

Those three have been visible, ID'd and age-stamped for eleven weeks.

Not lost. Not forgotten. Not buried in a comment thread on a ticket nobody follows. Listed, numbered, dated and printed in front of people every single week for eleven weeks.

That is the part worth sitting with. The register solved the problem it was built to solve, and the suites still are not written.

Every Tracker Encodes a Theory

This is the mechanism, and it generalises well past question registers.

A tracker contains a silent theory about why the thing it tracks does not happen. Build a register of open questions and you have asserted this: questions stay open because nobody can see them.

That theory is testable. Eleven weeks is a generous window. The questions were seen — by the same people, in the same meeting, with the age climbing in front of them — and nothing followed.

Visibility was never the blocker. Ownership was.

A register converts an invisible assumption into a visible one and then stops. That conversion is a real service; an invisible assumption cannot even be argued about. But conversion is not resolution. An age column is not pressure. It is a number that gets bigger while you watch it, and a number getting bigger is a measurement, not a mechanism.

The Two Fields That Turn a Record Into a Mechanism

A question row that can actually close carries two things an age column cannot supply.

• A named person, not a team. "The business" answers nothing, because "the business" cannot be chased, cannot be asked in a corridor and cannot feel a deadline. A team name in an owner column is a routing failure wearing a professional-sounding label. If you cannot put one human name on the row, that is the first finding, and it is a bigger one than the question itself.
• A forcing event with a date. Not a due date — a due date is another number that ages. A forcing event is something that happens on that date whether or not the answer arrives: a decision gets taken, a default is adopted, a suite is written against an assumption. The event is what converts silence into a cost for someone other than you.

Those are the only two columns I would fight for. Everything else in a register is bookkeeping.

Making It Work Without Authority

Most of us cannot assign owners. We can still supply the forcing event, and it is the half that does the work.

Write the answer you are going to assume, next to the question, as a default: unless this is answered by Friday, I will test it as X. Then test it as X.

That single sentence changes the shape of the situation. The question stops being a request for information, which anyone can defer at no cost, and becomes a decision that has already been taken by the wrong person. People who will happily let a question age for 73 days will not let a tester quietly decide their pricing rule for them.

It also converts a blocked suite into a written one with a stated assumption, which is a far better artefact than an empty suite and a complaint.

Why This Escapes Every Review

Because the register reports beautifully on itself.

It regenerates. It is complete. It is accurate to the day. Every quality property it was designed to have, it has, and anyone who inspects it will confirm that it is in good health.

The one number it does not carry is the number that matters: how many entries have closed since the last run. Age tells you how long you have been failing. Closure rate tells you whether you are failing right now. A register that only ages is a very precise record of your own drift, and precision reads as control.

That is the trap. 62, 44, 70-73, eleven weeks — those figures feel like a grip on the problem. They are the problem, stated in high resolution.

What I Would Add Before the Next Regeneration

One column for a named human. One column for what happens if the date passes unanswered. One line at the top of the health check reporting closures since the last run, even when the answer is zero — especially when the answer is zero, because a zero printed in a summary is an event and an age silently incrementing is not.

If your register has never printed a closed count, you do not currently know whether it has ever closed anything.

How old is the oldest question in your register, and who — by name — is going to answer it?