Ended up being easy enough, (while taking some shortcuts)
The polys don't actually roll, they scale down to 0 very quickly after flipping.
I modified the POLY compond to only scale in one axis or in length, for them to not also scale down in width.
I know it's very possible (even fairly simple?) to actually rotate matricies, you can check out some of "Probiners" videos for that..
but this can also work-out pretty well.
I also got a step closer and have the vectors back, attached a scene.
Problem is, that the force will push the vectors global, even when I change the compound inside to local.
So at one position the vectors are perfect and as soon as the NULL drives around the Sphere, the vectors become wrong.
I will play with your scene now, maybe I really have to get back to Polys
here is a little test btw.
the right and left side do not look so nice because of the orientation, thats the reason why I want to rotate away from null.
the grow effect is still spherical, but will change that to more irregular shape by weightmap grow.
The attached video is actually more or less, what I need.
Its only one pass. I use the same values to modify the domino stones, to control the opacity of the Earth.
Of course its still full of ugly artefacts, wrong spreading, wrong color, etc. But now I have to change it to the ICE particle grow like in that video anyway since they want it like that
So not only from one point but from many.
And I think, I than can use the neighbouring particles to generate velocity (there was a thread, where Paul explains how to do that, its something like subtract PointPosition from previous one, maybe I can rebuild that on this system) and align the particles to its velocity.
If From one point, I'm pretty sure you could easily orient particles towards a point.
I'll try something else maybe an evening later this week with one of your particle setups (hope your'e not too much in a hurry.)
I wish probiner could share some of his stuff. (not complaining he already contributed alot!)
Perhaps this might be of help, you can watch it in HD and maybe get glimpses of node setups.
He also uses Houdini...
I'm sure Houdini is great if you're very technical (like probiner) and have lots of time to do things,
otherwise I personally find Houdini -way- (way!) too extra convoluted for practically and literally everything.
That are some extreme sweet setups, so crazy stuff...
Ill check if I can sneak a code snip in his videos.
ICE modeling is just so cool and I guess, when you know it, its very quick.
I am sometimes playing around with houdini, too. Trying to do simple ICEStuff there.
Many things are genius, I like the node workflow alot. But the problems already start with different names for the same things.
One of the advanced things in Houdini, beside all the effect stuff - its so fast when you compare it to SI.
Just opening a massive scene with alot of geometry feels in Houdini so fluid while Softimage sometimes needs many minutes to open an equal amount of geometry.
But like you said, its quite a heavy software and when you are in love with Softimage, its more easy to do most things there.
Over the holidays we planed, to make skype sessions with workmates and learn a bit houdini together.
Anyway, I am again a little bit closer to my goal: it was the "Point Force" instead of "Null controlling Force".
I attached the scene. Super simple and I can either push away, or attrackt.
Now last step should be something like add the vectors from the "Point Force" to "Align Particles to Surface".
I'd be very happy if you try when you have time, just to understand it.
What I have a hard time most is scene management or hierachy representations.. and shading / making custom shaders, setting-up AOVs, .. and a bunch of things i'd do in ICE in a jiffy.
Most of all, I'd say is overall clarity of what does what exactly in varous setups, which is also related to my noobishness, but even when grinding down to figure out the where and why that makes a particular setup (any that doesn't heavily rely on VEX which is yet another level), I cringe at how unclear influences and dependencies are.
I haven't done any setups with constraints yet, but I looked into it, and for similar reasons, I don't want to even bother trying.
At least modeling became less bad or more direct in recent versions..
Also you can apply an "align to emit location" after (or before?) the "Align to position and Tilt" for them to stand, while oriented towards null
and note that doing so can make like a rift along where the threshold between two axis happens,
but can be dissimulated by moving the null around (less noticeable at different spots)
(could also very well be a way to completely address that 'rift')