I think it’s helpful to know if there’s a nonzero geometry issue, even when it is below the threshold. It could make sense to define an epsilon to account for numerical issues, but it can be hard to pick a valid value for that so letting the user look at the value and, as in this case, ignore it because it’s small.
To put it another way, the Examine step should probably generate some kind of “report” or “assessment” data structure that would describe how well the data matches the requirements of the reader. The user could choose to look at this or not.
Currently it’s true that the 0.001 mm threshold is very arbitrary and not exposed to the user.