Does the picture change the judgement? · Essai 03

Enough for the next question

A summary can preserve the correct answer and still lose something the next person needs. The test is what we expect that person to do with it.

The first essai gives six fictional handovers the same “4/4 preparations complete” summary. Their records nevertheless warrant different decisions. Completion alone cannot tell us whether both commissioning bookings may proceed.

Suppose we improve the summary. Instead of showing progress, we display the correctly computed verdict: Yes, No or Unresolved. Have we now kept enough?

Two correct refusals

Consider two of those existing handovers. All preparations are complete in both.

  Glasshouse pumps · C Hilltop relay · E
Consent Approved Refused
First booking Crew Fern, 09:00–11:00 Crew Ivy, 09:00–11:00
Second booking Crew Fern, 10:00–12:00 Crew Juniper, 10:00–12:00
Capacity One booking at a time per crew One booking at a time per crew
May both proceed as listed? No No

The verdict is correct in each case. In Glasshouse, Crew Fern is assigned to two simultaneous bookings from 10:00 to 11:00. In Hilltop, the bookings use different crews, so their overlapping times do not breach capacity. Consent is the obstacle there.

For someone who needs only the already-established authorisation verdict, “No” faithfully reports the answer in both cases. That assumes it was correctly obtained from the applicable record; the label cannot verify itself.

Now the incoming manager asks: “Is crew capacity what prevents this plan proceeding?”

The answers are different. A display containing only “No” to the original question cannot recover them. The two records have become indistinguishable at exactly the point where the new question needs their difference.

Enough changes with the use

Nothing in this comparison changes the facts or makes either original verdict wrong. The intended use has changed: from reporting the decision to understanding its grounds.

For this pair, retaining the consent state and the presence or absence of a crew clash would preserve the distinction. To check the clash independently, the recipient also needs crew identities, booking intervals and the capacity rule. To revisit the decision later, they would need to know which record the verdict describes and whether it is still applicable.

There is no instruction here to substitute a crew, move a booking or override consent. The original cases expressly exclude changes to the plan. Knowing why a decision is blocked does not establish authority to change its conditions.

This gives “what must a representation preserve?” a practical meaning. First name the next question. Then ask whether two records could look identical in the proposed summary while requiring different answers. If they could, the summary has lost a distinction that matters for that use.

The picture has another job

Keeping a distinction available is one requirement. Making it easy to notice and use is another. A complete table may contain the crew and time facts, while an aligned diagram places the bookings against a shared time axis.

The second essai tested that form contrast in four fresh model sessions using screenshots. Both the table and the aligned view produced 16/16 correct decisions. That result supplies no observed accuracy advantage for the diagram; it does not settle which form would help other readers or questions.

This essai is a worked interpretation of the existing cases, not a new reader trial. Its narrower conclusion is that preserving a verdict does not necessarily preserve its explanation, its checkability or its usefulness for the next question.

A useful handover question is therefore: what will the next person need to ask of this record? Keep the distinctions that answer that, and make their grounds inspectable.

Case evidence

The comparison uses cases C and E from the first essai’s case records, with their unchanged computed results. The methods specify the authorisation rule, fixed-plan constraint and independent checks across the constructed state space. No new observations or corrective-action permissions are claimed.