Equality problem

Replies to This Discussion

Permalink Reply by Joseph Oster on September 28, 2015 at 1:36pm

Try 'Similar' instead of "="?

Permalink Reply by Daniel González Abalde on September 28, 2015 at 1:47pm

You see on the panel 0.0, but actually may be eg 0.00000618, that's not 0. So use "smaller than" with a value as 0,001 instead.

Permalink Reply by Joseph Oster on September 28, 2015 at 2:06pm

That's what the 'Similar' component does.

Permalink Reply by Nenov Ivo on September 28, 2015 at 2:15pm

Thank you very much !!

With smaller than it's working but not with similarity, it's strange...

Permalink Reply by Joseph Oster on September 28, 2015 at 2:19pm

Did you try adjusting the 'T%' value?  The default is 0.1 percent.

Did you find out what the actual value is?  And why it's not zero?

Permalink Reply by Mateusz Zwierzycki on September 28, 2015 at 2:33pm

http://www.grasshopper3d.com/forum/topics/floating-point-error-why-...

Permalink Reply by Nenov Ivo on September 28, 2015 at 3:08pm

Ok thnks, after reading this i kinda understand why i got this problem...

Joseph :I try with the similarity, it's still not working

            I don't know how to find the real value... 

And even with this number problem solved i didn't managed to succeed the tutorial... ???

Sometimes easy things seem so complex... maybe it's because of my fever hehe

Permalink Reply by Nenov Ivo on September 28, 2015 at 3:29pm

Here's the code. It will be very nice if someone can check the code and explain me why it's not working ? I change the values, change the surface... nothing goes like it should !

Permalink Reply by Nenov Ivo on September 28, 2015 at 3:30pm

code

Attachments:

• PlanerizeHexa.gh, 11 KB

Permalink Reply by peter fotiadis on September 29, 2015 at 6:52am

Hmm ... I say

1. This is WIP (with regard closed surfaces: read commends AND DO NOT use the closed example).

2. Relaxing "patterns" in general within a "boundary" is a bit tricky because requires a rather fine balance between goals ... er ... pardon me ... "constrain classes" as they are listed in the def. It could be way easier to relax by pulling against an extended surface (meaning no restrictive boundaries). Spot the yellow group dealing with what kind of points pull to the outer Loop (a joined curve in this case). Make this inactive and test the result.

3. On the other hand ... If you vary even slightly the K values > Armageddon ... that's why the next update (V2) uses the preset concept to guide you "safely" into some meaningful result.

4. C# is used in order to extract "whole" hexagons from the rest. But actually the reason for using this approach via code would be clear in the forthcoming V2 (where lot's and lot's of patterns are available made internally - without  Lunchbox).

BTW: do you use K2? have you experienced benefits using the multi threading optional capability?  If yes on what O/S, CPU?

BTW: Imagine Daniel getting his hands dirty with CUDA...

best, Peter

Attachments:

• Relax_Hexagons_V1.gh, 170 KB

Permalink Reply by peter fotiadis on September 29, 2015 at 7:00am

For instance (no boundary restrictions):

Attachments:

• K1_Relax_panels_on_nurbs_and_measure_14B.gh, 111 KB

Permalink Reply by Nenov Ivo on September 29, 2015 at 11:00am

Thank you very much Peter for your answer !

At first sight, i thought i'll never understand your code but it's very nice and clear.

I remake it without C# things (they make me feel dumb hehe) to understand it completly and now I'm wondering: if i let kangaroo goes for a long time, will they reach a planarity or they will always be almost planar ?

BTW i've never used K2, i tried to install it some month ago but it didn't worked, i have to retry...

Checking the second file when i come back from my evening lecture. Thanks again !

Attachments:

• Relax_Hexagons_V1noC.gh, 75 KB

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