# Pull to unpredicted results, non-unique points

… don t want to call it bug, because this is (maybe) more a geometric problem.
… and we still don t have the category modelling / geometric understanding

pulling some curves to a surface, i got really strange results - after further investigation i could somehow reconstruct this excerpt.
(came across it, modelling a quite geometrical chair)

It is obvious, that the base/center-point of a cone does not have a unique solution being pulled to a con. see green example in the middle, all points on the green circle have same (shortest) distance to the center.

but also pulling a curve to other geometry - gives results that are not geometrically correct.
see shelter-like surface in the background.

cone_pull.3dm (3.7 MB)

for me it seams, like the pull command does not handle those special cases in a meaningfull way.
it should be possible that a curve that get s pull to a surfaces has more than one result, with a gap (those two small curves in the background.)
and for some special cases pulling a point to a surface - the correct answer is a curve.
(there is other examples, where a geometric operation has different types as result: a intersection between two curves can have an overlap or a point…)

just came across this, and wanted to share it.
i would guess @theoutside and hopefully other love this geometric excursion ;-D

kind regards -tom

1 Like

Hi Tom - it looks like at a certain point samples on the pulled curve - points - are ignored as being too far off or some recursion limit is reached… I don’t know what the inner works are doing… so a subset of pulled points is used as input. You can see this on the cone/line if you divide the line into say 32 points and pull these. The point near the center of course will make a mess of things too.

-Pascal