2D Curve scan troubles

[[Pe...]]

Reversing the scan... you mean actually running right to left instead of left to right? Or do you mean add a second segment and inverse it as to have it run both ways?

[[Ro...]]

Either. Running the opposite direction can help you determine if any errors you're seeing are actually on the part or a byproduct of the scanning process.
Example... If you have a spike at one end of a curve and it isn't there when you scan the opposite direction, you know that it's probably related to the machine movement and can be filtered out. It's just another piece of information that could help you.
Another thing I try and do is always start my scan on a flat surface if one is available. I will cut and paste points to get my start point in the middle of a flat section. This helps the probe to settle and get started on a surface that has one direction vector, as opposed to starting on a radius and having the machine ramp up and change vectors at the same time.