Trying to write the lathe handover in words our operators can use. Machine supplier says ready when the request arrives. Integrator means loading area available. Both keep saying they agree, and my draft now contains one word with two quite different moments hiding behind it.
Give the moments different names before choosing signal labels. Ask the machine side to describe receipt while the machine is still busy, then ask the integrator what it would wait for in that condition.
Machine side clears its receipt indication on a timer. Integrator assumes ready remains available until transfer begins. We've put those statements side by side; they describe different behaviour as well as different events.
That comparison should include the receiver's response to loss of the indication. Otherwise changing the producer's duration may simply conceal the mismatch during the normal sequence, while leaving the transfer decision undefined when the condition changes.
I'd walk a paper example through receipt, still busy, then unavailable again before transfer; cheaper than arguing over a signal abbreviation for another meeting.
Elliot, make the late receipt belong to its original request too. I've seen timeout discussions stop at failed, then a late answer quietly gets used for the next request because nobody specified the relationship.
Owen, the machine controls lead owns that condition definition. Harish, the cell lead owns the recovery explanation. Elliot and Hazel, we've added the busy and late-receipt examples, with request identity visible in the paper sequence.
Receipt has an agreed identity and clear rule now. Availability has a machine-side condition definition, but the integrator still wants a separate explanation of what happens if it is withdrawn while a transfer is being prepared.
That is a reasonable unresolved case. They need to distinguish preparation before transfer from a transfer already in progress, using the observations the design actually provides. Neither response should be inferred from the signal being called availability.
Yes. They jointly own the open interrupted-state entry, and the cell lead will review the operator response with them. Receipt and normal availability are documented separately; I haven't marked the complete interface agreed.
At least the missing decision now has a name and owners; keep the awkward paper example with it so nobody closes the entry by demonstrating only the easy sequence.