The Missing Shipments: A Robert Johnson Investigation
Copyright© 2026 by Alexander Winston
Chapter 5: One Witness Wearing Three Names
Rita moved him out of the control room before he started. The preserved export was a copy, and a copy could be read anywhere. She set him up in the engineering office on the ground floor with the door shut.
“Whoever can reach that gateway works in that room or walks through it,” she said. “I’m going back up to ask for more reconciled reports. I’d like us to go on looking like two auditors checking the arithmetic.”
He took one shipment, the lenses, and gave it a page to itself. Three columns: what the record claimed, which system held it, and where that system had learned it. The third column was empty.
MERLIN, the routing manifest shows MK-45066 dispatched from Dock Three at 02:14. What event caused that entry to be written?
Routing wrote the dispatch entry on receipt of loading-complete event LG-7718203.
And the inventory deduction?
Inventory deducted forty cases on receipt of loading-complete event LG-7718203.
Robert looked at the two numbers for a while.
The dock activity report shows a loading event at Dock Three at 02:14. That one comes from the door sensors on Dock Three, doesn’t it?
Partly. Door sensors report open and closed states to the loading-sensor interface. The interface associates those states with cargo scans and issues a normalized loading sequence. The dock activity report stores the normalized sequence. The Dock Three entry at 02:14 derives from event LG-7718203.
What issues events with the LG prefix?
The loading-sensor interface. Staff call it the scanner gateway.
He filled in the third column. The same line, three times.
It might mean nothing. Perhaps everything in the building ran through the gateway and that was simply how it was built. He picked an ordinary shipment from the same night, a load of garden tools that had reached its consignee without complaint, and asked the same questions. The answers had the same shape: one gateway event, three downstream records. So the routing was normal. That was the trouble with it. On any night at all, the three bookkeepers he had been so pleased with were copying from a single sheet. They kept separate books. They had one informant.
MERLIN, is there any record of MK-45066’s physical position that did not pass through the gateway?
Not among the records I use. The scanner heads at each dock retain a local diagnostic buffer for fault analysis. It is not part of the reporting feed, and I do not read it. Since this morning’s preservation instruction those buffers have been exported. I can list the export references for Dock Three and Dock Six for the relevant period.
Please.
The diagnostics were ugly, engineer’s data: head temperatures, read counts, signal strengths, nothing meant for a manager to see. That was their value. Between two and half past two on the Wednesday morning the scanner head at Dock Three had logged no reads at all. The gateway said forty cases had passed under it. The head at Dock Six, in the same half hour, had logged forty-one successful reads, and the gateway had reported none of them.
Robert wrote: Device diagnostics inconsistent with gateway output for both docks. Diagnostics do not record cargo identity. Forty-one reads were not forty cases of lenses. They were a reason to look inside the gateway.
The preserved configuration ran to some thousands of lines. Most of it was what he would have expected: device addresses, timing tolerances, the rules for turning a messy burst of scans into a clean sequence. He read it through from the top because he didn’t yet know what he was looking for, and near the end of the third hour he found a block filed under maintenance transformations, between a rule that corrected a mislabeled scanner and a rule that ignored a test tag.
To read the complete story you need to be logged in:
Log In or
Register for a Free account
(Why register?)
* Allows you 3 stories to read in 24 hours.