The discrepancy of one

People say there's no such thing as progress, but I'm here to tell you that there is. How else to account for the fact that what began as a discrepancy of four and was, less than 24 hours ago, still a discrepancy of three, turned last night, while I was completing the recount already under way when I reported on this, into a discrepancy of one? The most interested among you might be wondering how the previously overlooked items get to show up. I'll tell you: they are - the rediscovered three anyway - extremely slender publications whose spines I've failed to notice during the first count because they are too far in on the shelf, hidden between two other books.

A reader emails me as follows:

Dear Norm, As a librarian I can certainly sympathise. You have hit upon a sixth law of librarianship, spoken of only in the dark corners of library conferences and between the rolling stacks, and never in front of non-initiates.

This is that the catalogue and accuracy are close relations, but not identical twins.

Still, the joy is in the reading and not the cataloguing!

Grateful as I am for the relaxed approach encouraged by the 'close relations/identical twins' distinction, the sharp observer will have realized I still have a problem. A discrepancy of one remains a discrepancy. There are some who wouldn't worry about it. I count myself one of their number. But you've got to admit, it's intriguing. To what can it be due? (Thanks: BT.)