Indeed it assumes the stamps are scanned upon sale. Otherwise, at best, what you obtain is "where was this stamp sold", which is completely non-identifying. Bu

Indeed it assumes the stamps are scanned upon sale. Otherwise, at best, what you obtain is "where was this stamp sold", which is completely non-identifying. But if you think about it for more than ~10 seconds, you remember that stamps are sold in packs of 4-8 at least. They're also often sold in sheets of 50+. All this makes individual scans painful. So you want to play the devil's advocate and make more assumptions; say, for example, "every stamp could trace back to its original pack, so only one of them needs to be scanned". But if that's the point of this panopticon move, then why put the barcode on the stamp, rather than … the pack itself? Because you know what always gets scanned? The product's barcode. If I wanted to track people who buy stamps, I would create unique EAN13-compatible barcodes for every single pack sold, always start them with the same number space + some digits for actual different products, and voila, somewhat backwards-compatible, unique tracking on all those stamps without having to deploy a barcode on every single stamp, implement a swap-out scheme, etc. Of course, maybe someone at GCHQ is reading my comment and thinking "fuck, why did we spend so much money on this scheme again?", but somehow I doubt it.