Distance Measurement Point3d vs. Point3f

Grasshopper
, , ,
1

Hi,

I am having some tolerance issues when measuring between points/ vertices and I wonder what the reason for it is. It seems like there is a discrepancy measuring between Point3f which is used internally for Mesh Vertices and Point3d. Please see attached screenshot as an example.

I am working on a facade and try to determine repetition of panels by comparing quadfaces of meshes. I measure for this all 4 sides of a quadface + the 2 diagonals to determine congruence. From measuring between the Point3f I get different results then from measuring my initial setout points. The error occurs at the 4th decimal - as I then round to the 3rd decimal it does affect my endresult.

Is this a problem of double vs. floats precision ?

By the way my doc. tolerance is 0.001

Many thanks,
Philipp

2

Wikipedia says floats can have between 6 to 9 significant decimal digits. Meaning that if you are working with something that measures over 100 meters, anything under millimeters might not be accurate.

3

Although from your description that might be a panel of a façade in cm. In this case your significant digits reach the microns.

Edit: If the panels are not stacked near the origin but placed in the buildings facade, your significant digits go back to mm as the coordinates used to calculate the edge lengths will contain much larger numbers.