it looks great, unfortunately have to switch to urugent work now, will check in detail later! thanks!
Hello, I tried it, but having some imprecise results, only few objects changed:(
not sure it is curve the resampling issue when objects are varying in the widths too much.
-without the named attribute it is a lot better, why do you use it?
-is there way (in the repeat zone, sampling in all curves mode, I have info from points before the position is applied, I just divide the spline lenght by largest accumulated mix value with added 1/2last instance random scaled width then this multiplier using to correct the accumulated mix value right before set position node) correction multiplier = L/(acmmax+last1/2iw) this aligns all instances nicely from first point of the spline to the last.
here I am trying to sample these values and do something similar, but no success so far to acces the total of the accumulated field per spline yet… (tried couple ways but with the viewer I am getting similar but not exactly right values)
Whups… I should not have taken in sorted index (red line) - fixed now uses the unsorted one (cyan one):
Here is fixed file:
Pack curve1.blend (116.1 KB)
Huh? I get exactly the same result… I did it so the network can be easier to understand as it “labels” the attribute and you can see it in the spreadsheet more easily. You can totally replace it with a capture, it should not make any difference. I added a switch in file above to prove the point. Take the named attribute out if you want.
Edit: no more capture… ok?!
I even have not used the capture I connected the value straight and that worked…
Oh right - that is a good point. The sample Index effectively “captures” it so you’re right. It doesn’t need it… Initially it was just there so I could see what I was doing in the spreadsheet.
File updated in previous post.
I am not right, it was just try, I still half understand to what I do…
thanks and if you know how I can read the last accumulated instance value per spline
As to your other question you just added…
Accumulator Total is a lie, so better to use the Curve Parameter Length property of the last point… delete all excess points and you should be able to use the end-point selector. (Never convert to points. keep the “points” in spline form for the whole operation and this should be easier.)
Then do a 2nd resample where you add half the excess to the first point.
Something like that… 
Good luck.
Hey, sorry, I have no idea what I’m looking at?
Hmm… right, so the indices are changing meaning the random value is changing which means the values will change… yeah, always trickier than you suspect…
Maybe try something like this on the index selector to ensure the random values will be the same…
Good luck.
awesome, thanks, I was trying in that section as well, but connected to the ID socket index evaluated on spline domain only and wow that populated each spline with singe index instance…
finally best results in my testing are with simply using the total value…;/
z-pack-curve-03.blend (188.0 KB)
