The application clears it on its next inspection. The contractor's proposed PLC change would also clear it. Neither document describes a restart.

Two writers for one state is a poor starting point. Agree the sequence first. Include a new request arriving while the old receipt is still there.

I had a handover where both teams said their bit was only a mirror, so a table naming the actual writer of each value saved more time than another screenshot.

I have asked both authors to fill in that table. The application author also says received means accepted for inspection, not inspection finished.

Does the operator display make that distinction?

Felix, a retained receipt can be useful if its identity is clear; I would not insist that every receipt vanish immediately just to make the display tidy.

I didn't ask for immediate clearing. I asked what happens when the next request meets it. Retaining a receipt with no matching rule is the problem here.

Anil, the display says received, with finished separate. Both authors have now agreed that the PLC owns request and the application owns receipt. Their draft matches an attempt, withdraws that request after receipt, then has the application clear its receipt when it sees the withdrawal.

And what prevents a new request being issued before that clearing is observed?

Put that transition into the same drawing, Oscar; otherwise the two authors can each assume the other one supplies the wait.

The PLC author has added that wait. We are using the diagram to build the offline cases, including a receipt that arrives after timeout.

For the timeout case, keep the uncertain attempt available to reconcile. A fresh identity should not quietly turn an unconfirmed physical sample into a different sample that everybody assumes is untouched.

Has either restart case been run yet?

Application restart was run. It wrongly cleared the outstanding receipt before reading the retained request. PLC restart is still waiting its turn.

That is worth finding before commissioning. Who is defining how the uncertain sample is reconciled after that restart?

Our integration lead and laboratory lead own that response together. They have not finished it. I have authorised the next offline test work against the agreed sequence, not an installed change; at least the quote now names what we need answered.