> On the contrary, it seems totally logical. If the difference is so small that you can't assuredly decide between two options, either will do. A slightly diff

> On the contrary, it seems totally logical. If the difference is so small that you can't assuredly decide between two options, either will do. A slightly different way to interpret this would be to say that that particular axis of variation is at its limit, given the state of the other axes. There are several axes, and they aren't perfectly orthogonal. Zeroing in on one can alter the best correction achievable on the other. Not all opticians will go back and iterate other axes, but I think it's a good idea. Lock X, find Y, recheck X given Y, etc.