There was a science fiction writer some years ago named James Blish who took the first Star Trek series characters and wrote variants of the first series' episodes. Sometimes he added amusing additional dialog between Mister Spock and Doctor McCoy. One such exchange was about the transporter.
The premise in James Blish's work was that the transporter worked by recording every last particle of the transportee (forgive my coinage of that word) down to the sub-atomic level and then reconstructing the transportee at the transporter destination.
This bothered Doctor McCoy very much. He felt that the real Doctor McCoy had died many years ago when he stepped onto a transporter platform for the first time and that he himself was now only a duplicate, not the actual person. To Mister Spock, this was of no consequence at all. Since the reconstruction was absolutely perfect, there was nothing about which to voice any objection. "A difference which makes no difference is no difference." he opined.
Of course, this philosophical disagreement was never resolved between them, but the argument may have some applicability in the real world. Please see the following images, the original one on the left and the modified one on the right.

For no particular reason one day, I ran my scanner over my wristwatch just to see what the result would look like. Then I noticed something. The angular steps of the second hand were not all of the same size. There were differences that I highlighted on the right image just to make them more visible. Here was Mister Spock's "difference which makes no difference".
Of course, this becomes a question of what are acceptable tolerances under some given circumstance. I have no difficulty with a substantial difference between angular step sizes of the second hand, but I do want the sum of all sixty of the step sizes to be extremely precise!
So does this have any bearing on proper development of error budgets?