skeleton

Comment

Comment by martyn hogg on February 16, 2014 at 4:26pm

Infinite Monkey strikes again... Reversing the list of points worked!

Comment by martyn hogg on February 16, 2014 at 12:21pm

Hi Michael, I tried reconnecting but it doesn't work for me... could be the dreaded Rhino4 instead of Rhino5 bug :)

If my hexagons intersect the 3dHulls often fail completely... the size and distance from the node of the hexagons is dependent on the angle of the lines... acute angles need the hexagons moving further away to avoid them intersecting.

Comment by martyn hogg on February 16, 2014 at 11:20am

a similar effect can happen if the hexagons are not removed from the 3dHulls

Comment by Michael Pryor on February 15, 2014 at 6:56pm

That is because your hexagons are intersecting. I've come up with a way to avoid this i will release shortly. Basically per node your hexagons need a rule to not be bigger than half the distance to its neighboring hexagons in that node.

Comment by Cristina C on February 15, 2014 at 4:09pm

Hey guys. I'm having trouble with the joints between the nodes and the pipes. Some of them seem to not be blending properly as you can see in the picture.

I'm using sHull3d for the pipes also, and the way im organizing my lists for them is still a bit sketchy, I'm not sure if this is what causes the problem.

You can find my definition here. And here is my rhino file.

Any tips?

Comment by Michael Pryor on February 9, 2014 at 1:13am

martyn I am als0 get a red hull depending on my points but I just re-connect the wire to the hull and it turns grey again. Have you tried that when it turns red?

Comment by Cristina C on February 7, 2014 at 5:39pm

That works great Michael, definitely more precise than what I was doing before. Thanks

Comment by martyn hogg on February 7, 2014 at 2:08pm

I path Mappered the data to {0;0;0} but still no luck so maybe I'll ask Mateusz!

Comment by martyn hogg on February 7, 2014 at 10:36am

Thanks for that Michael, that makes sense!

Comment by Michael Pryor on February 7, 2014 at 8:47am

Martyn I am not sure I suspect it may be that your one branch is 0;0;1  which maybe throws of the component as a branch list should I think start at 0;0;0. You may have to trouble shoot this with Mateusz the maker of starling.  

Cristina and Martyn, for the nodes I found a really good way which is less accident prone then the less than method. What i do is evaluate each cure at length .01 from both ends with evaluate length (not evaluate t). Then I measure from node center to these points and if the distance = .01 it belongs to that node. Then I make a vector from node to eval points and move them outward then 3dhull to make the node size. The reason for this is with less than method you may have an eval point that you want to belong to one node but based on your node network may actually be closer to another which wont result good. The equals .01 method basically ensures the eval point is always with the correct node.  This method will also let you enable varying node sizes base on a list of numbers or attractor as you are assigning the distance (node size) after you have grouped them as a node. 

• ‹ Previous
• 1
• 2
• 3
• Next ›
• Page