I have built this in GH, but it is a Rhino question that pertains to surfacing.

Divide the top and bottom cirlces of a pipe and connect two points across the pipe surface. Simple right? as it turns out, no.

In a surfacing video, it was suggested that PULL is the optimum solution… right?

In these threads the question was asked, what are you trying to achieve?


The goal here is to blend these surfaces together such that tangency is maintained along the length of their intersection. Tangency is set at the top and the bottom, but each of these methods results in a slightly different path between, which will show up when the they are merged and may result in a naked edge.

This is perhaps the biggest problem I have had in surfacing so far.

I have built this example in GH so all the steps can be analyzed without making a click by click video. I have chosen a planar pipe such that the two points to be connected are equidistant from the XZ plane so the midpoint can be analyized more easily.


Project is not optimal becuase the plane of the projection should be normal to the surface which it cannot be for this pipe. i created the surface in the Y dir to optimize PROJECT


The curve must travel across the surface of the pipe

tangency at the midpoint is known



If a plane bisects all the lines at the pipe vertical midpoint the differences are clear

Since all three methods are correct in their own right, which one is consitently tangentially correct along the length of the curve ?

SurfaceCurve (27.0 KB)

Hello -

what are the differences, exactly?


. deleted

Hello - those minute numbers are reported in “scientific e notation” and are pretty close to zero - What I see is 5.6205 e-8 = 0.000000056205



Pascal answered a very similar query I had some time ago. Try Shortpath between the top and bottom points.

well thats embarrasing. How did i miss that…

Try Shortpath between the top and bottom points.

Thank you for that. Another new command…