Unless, relatively speaking, you get a precision tape measure. The survey tapes were always labeled with calibration date and temperature compensation data.
On further reflection (& close inspection of my ruler) I think there may be another explanation, which has not been mentioned in this thread.
In the case of metal rulers, and possibly plastic ones as well, if they ended in a perfect 90 degree angle corner, it would be a sharp corner and liable to scratch people. So the ends tend to be rounded. However, you also want the ruler to have 12 inches of perfectly straight length, so you need to start the rounding after the beginning and end of the 12 inch span. As a result, the full length of the ruler will inevitably be more than 12 inches. Looking closely at my ruler, ISTM that the rounding begins at precisely the 0 and 12 inch marks (& reaches 90 degrees at ~3/16 of an inch).
When I saw this thread topic I thought it was “rulers” as in “leaders” and thought this was going to be about something else entirely. 
0.050 inch as well. For integrated circuits, DIP packages (still widely available, and nice for prototype work) have 0.1 in lead spacing (pitch). SOIC is 0.050. TSSOP is close to 0.025 in, but they defined it as 0.65 mm instead. For small packages you could get away with snapping to 0.025 in.
Surface-mount components tend to be listed in imperial as well. For example, a “0402” surface-mount resistor is 0.04x0.02 in. Though they can be labeled in metric as well, so a 1005 represents the same resistor, 1.0x0.5 mm (the small error doesn’t make a difference here). It’s somewhat confusing since the codes don’t tell you which system they’re using, and they’re only mostly distinct.
Could be worse. I was running an instructional lab once, and once group kept on getting results that were slightly, but consistently, off. It turned out that their “metric ruler” they were using was marked, not in centimeters, but in 32nds of a foot: Close enough that you’d never notice without a direct comparison or reading the fine print, but far enough off for noticeable differences in the numbers.
Sorry, that should have been “~3/32 of an inch”.