I was puzzled recently by what I've read in the internets about Heisenberg's uncertainty principle (one probably should never do this).
It claimed that in the usual relation $$ \overline{(x-\overline{x})^2} \cdot \overline{(p_x-\overline{p_x})^2} \ge \hbar/2 $$ uncertainties $\overline{(x-\overline{x})^2}$ and $\overline{(p_x-\overline{p_x})^2}$ correspond to the limits on precision of measuring instrument (e.g. we cannot build a device that would measure the momentum precisely).
But I always thought that the meaning of the above relation is as follows:
We have a large number $2N$ of "boxes" with identical systems, all in identical quantum state. We then measure coordinate of $N$ of them and momentum of the other half. We can make each of these measurements, both of coordinate and of momentum at the same time, as precise as possible, with instrumental error virtually equal to zero.
We then find that despite being in the same quantum state, results are not the same for different boxes. And after averaging over boxes $$ \overline{x} = \frac1N \sum_{box=1}^N x_{box} $$ $$ \overline{p} = \frac1N \sum_{box=N+1}^{2N} p_{box} $$ etc, we find the uncertainty relation. So each component in this sum is a precise value (in principle).
Is my understanding correct?