A registered item is signed for on a pad at the counter and the signature is stored with the item's barcode and the time, and an item with no signature is an item that has not been delivered. Every night the pad stores a signature for a barcode that is not in the building, was never scanned in, and belongs to a batch of labels issued in 1998, and the signature is legible, and it is the same name every time.
I am the counter supervisor, and the signature pad's records are what I produce when somebody says they never received it.
How the pad works.
An item is scanned, the customer signs on the pad, and the strokes are captured with the barcode and the time and the counter position. The record is kept for years. A registered item that reaches its destination has a signature record, and one that does not has a trace that ends somewhere, and my job is to find where.
The pad is at the counter. The counter is closed at half past five and the office is locked at seven.
Every night at about half past eleven the pad records a signature.
The barcode with it is not one we hold. It is in the right format and its prefix is from a batch of registered labels issued in 1998 and used up by 1999, and the national system has that number as delivered, in 1998, with a signature record, which I have pulled, and which is the same signature as the one the pad captures every night.
The signature is a name. Legible and careful, a taught signature from a time when signatures were taught. A woman's name, and I have looked her up in the way the office lets me, and she was the recipient of that item in 1998, at an address in the town, and she signed for it, and the signature on file from 1998 is hers, and the signature the pad captures at half past eleven is hers, the same, stroke for stroke.
What has been checked.
The pad, replaced twice. The counter system, updated. The network, which the pad uses to send its record, and which shows the record arriving from the pad at half past eleven from a counter position that is dark and locked.
The camera over the counter. It shows the pad, dark, and nobody at it, and at half past eleven the pad's screen lights, the way it does when a pen touches it, and stays lit for the time it takes to sign a name, and goes dark.
The pen. It is on a cord. On the camera it lifts, and moves, and lays down, and there is nobody holding it.
The item.
I have the 1998 record. A registered letter, from a solicitor, to the woman, at her address, signed for on the twelfth of a month in October. The solicitor's firm still exists and I asked, without expecting an answer, and the clerk looked it up and said it was a letter about a will, and that she should not have told me that, and that the file noted the recipient had died the following week.
What is done about it.
The record is stored, as every signature is, and it shows up on no report, because it is not attached to a live item, and the national system, which I have queried, has the barcode as delivered in 1998 and does not accept a second signature for a delivered item, so the nightly one goes into our local archive and nowhere else.
There are about two thousand of them now.
What I do.
I supervise the counter.
The one thing I have changed is that I do not do the end of day on the counter's own screen.
The end of day used to be done at the counter, with the pad beside me, at half past five.
I do it in the back office now, and the reason is an evening in the autumn when I was late, and it was after eleven, and I was at the counter finishing, and the pad lit beside me. The pen lifted on its cord and wrote, and I watched the strokes go onto the screen, the careful hand, the name, and when it had finished the pad did what it does when a signature is captured, which is to show the item's details for confirmation, and the details it showed were not the 1998 item. They were a registered letter, from a solicitor, to me, at my address, with a barcode from this year's batch, and the pad asked me to confirm receipt, and I did not.