A mil, or a thou, really isn't good enough for high precision work. 25.4 microns is a lot. A surface flat to that isn't really that flat. An instrument that measures to that isn't really a very high precision instrument. Microns are a pretty good unit.

EDIT: There's this rule of thumb (I might be getting this slightly wrong) that one in a million accuracy/precision is about what you can expect to achieve without resulting to exotic means (at least order of magnitude). So microns of flatness over a meter e.g. Trying that with imperial units just breaks my mind, what's a million mils. Sure I can calculate that. But it's like a pretty useless measure. So accuracy of one thou over a million mils is sort of weird. (1000 inches which I'd have to translate to some normal unit... 25.4 meters ;) ).

Yeah, that's why tenths (~2.5 microns) are used when thousandths aren't precise enough.

I'm definitely not claiming it's a good, I'm just describing the system I have seen used in the handful of machine shops I've spent time in.

I think YZF just mis-parsed your sentence, and thought you meant a thou is for very high precision work, while you meant that a tenth is for very high precision work.

(That said, I'd add -- measuring to one part in 10^6, for anything except time, is challenging but still very doable; machining to one part in 10^6, for anything beyond very simple shapes, does require pretty exotic methods or very great care, and to verify it you have to measure to one part in 10^7).

Oh, yep, there's an extra comma that makes my comment a bit ambiguous. Alas, it's outside the edit window...