there's a lot of techniques used in such research which is not 'public knowledge' (even though a lot of techies might know them). Hence, information might purpo

there's a lot of techniques used in such research which is not 'public knowledge' (even though a lot of techies might know them). Hence, information might purposfully be a bit vague, to give the idea they did not reach certain conclusions. I'd expect they would, as you, know the exact model and make of the printer if it was microdots. They can also have libraries of things printed with lots of printers to analyse the quality of print etc., and then get an estimate for example by a neural network examining artefacts. in such a case, it would never be 100% certain, but maybe 95-99% range somewhere if they do it good. I can _imagine_ such techniques might even also get to the point where they can somewhat certainly identify other aspects of the pc like graphics cards, even though they don't know exactly how the network will draw these conclusions.