April 6, 2012 - Surface from points

Replies to This Discussion

Permalink Reply by Andrew Heumann on April 5, 2012 at 11:36pm

Hmm, I might be pushing the parameters limitation a little bit with these, you be the judge. Attached are 3 attempts.. only one works with the second point set. 

Attachments:

• Andrews_3_tries.gh, 34 KB

Permalink Reply by Yoann Mescam (Systemiq) on April 6, 2012 at 6:56am

I didn't want overusing of parameters to deal with the problem, so 1 more parameter is ok.

Your definitions are close to mine, which also uses the fact that the shape is convex to sort points.

(convex is a hint).

Permalink Reply by Andrew Heumann on April 6, 2012 at 7:16am

ahh, that was a big hint. 

Attachments:

• Andrew_try_4.gh, 9 KB

Permalink Reply by Yoann Mescam (Systemiq) on April 6, 2012 at 7:52am

Nice, let see if someone manages to get a shorter definition.

Mine is 21 components (including the 3 parameters), working on the second set.

Yours wins for the first set, good using of X to replace branches directly (although you can add a few more components to avoid this nasty shift).

Permalink Reply by Andrew Heumann on April 6, 2012 at 9:13am

15 components, strict about parameters, works on either point set.

Attachments:

• Andrew_try_5.gh, 13 KB

Permalink Reply by Yoann Mescam (Systemiq) on April 6, 2012 at 10:54am

Hehe, clever optimization, the plane rotation gives correctly oriented hulls, so you can use directly the indices of points on the hull.

Here was my solution.

Attachments:

• Ghafé2-Apr-6_Systemiq.gh, 12 KB

Permalink Reply by Andrew Heumann on April 6, 2012 at 11:06am

Ah nice, clever to use seam on the convex hull itself. 

Permalink Reply by Danny Boyes on April 6, 2012 at 12:36pm

Here's my wacky effort

Attachments:

• ghoffee_20120406_3DSort_Native.gh, 15 KB

Permalink Reply by Yoann Mescam (Systemiq) on April 6, 2012 at 3:57pm

Interesting approach !