Force vector on flexiprims: Why global only?

Thanks for the ability to specify a force vector on flexible objects! It really opens up a whole new set of possibilities.

One complaint I have, however, is that this is a global vector only. In other words, if a rotating object uses the force setting to bend in a particular direction, it will always bend in that absolute direction regardless of what angle it is rotated at. While this behavior may be desirable in some cases, there's plenty of others where the creator may want the prim to flex relative to its rotation.

Now, it would be possible to script such an object so that the force vector constantly changes based on the object's rotation. The problem is that this won't work for avatar attachments. The only worldspace-relative rotation detectable on an av is the overall rotation of the av itself, whereas the worldspace-relative rotations of attachments are affected by the av's pose/animation at any given time.

This means we can't currently make flexible hair strands, animal whiskers, etc. that consistently bend in a particular direction relative to the attachment.

If at all possible, I would really like to see an option added to the flex UI (and a corresponding parameter added to llSetPrimParams) that determines whether the force vector is applied globally or relative to the object's root prim (and possibly even a third selection for making it local to the prim itself).

Anyone else think this would be useful?

On the one hand, yes, it would be nice to have a local force option.

On the other hand, physically, the global force is correct for things like hair strands, whiskers, and so on. This gives you the only global orientation possible for attachments, which not only allows you to have avatar attachments that can move correctly with respect to external objects for the first time ever. Like my arm tentacles.

The problem is that flexible objects can't be "bent" objects, like truncated tapered toruses or cut slices of spheres. If you could apply flexi to existing attachments that would give you the best of all worlds.

So if there is only one force, it should be the global one, and if LL is goingto change something they should work in the direction of better coverage of flexible prims.

But there really should be a local force as well... best would be to have two force vectors rather than one that could be in either direction.

Err... I'm not quite sure what you mean by this. Hair strands and whiskers don't move with respect to external objects like your tentacles do (which are really cool, by the way).

Hair strands and whiskers tend to be curved in a certain direction, as opposed to simply being pulled straight down by gravity. This is where I pictured the local force coming in -- it would act to bend the prim in that direction. That wouldn't be doable with a global force, because the prim would always be attracted to that global direction whenever the person turned their head.

Again, that's what the local force would be used for: to bend flexiprims in a direction relative to the attachment. This would be done in place of using tapered torus segments.

I do agree that it would be nice to have other prim types be flex-capable, though.

That thought had crossed my mind, initially. I think I had dismissed it because it seemed overly complicated, and I couldn't really picture a case where using both was desirable. But I suppose that is a more versatile solution... you never know what ideas people will come up with!

No, but they're moved by acceleration and the wind of motion in ways that the "drag" doesn't really match. I've been playing with the idea of applying a force based on llGetVel to try and ge better bounce and loft.
Well, whiskers really aren't flexible on a scale that LL renders anyway. Hair is, but on that scale it really is just pulled by gravity. It's only "curved" because it interacts with the head. I don't think you could get near as good results with a local force as you could with stiff-ish toroidal flexiprims.

Have you tried scripting the force with llRot2Up(llGetRot())? The other llRot2* functions specify different axes and you'll need to put it in a timer to constantly update the rotation (since there is no-client side version--AHEM LL!).

You can't get the rotation of the avatar in-game, because as far as teh sim is concerned an avatar is always a grid-oriented prism. All the rotation and relative movement of attachments is handled by animations and that's entirely client-side.

No, Beatfox's problem's real, and either solution (his option to specify whether the force is local, or my local and global forces) would work, but without them there's really no workaround you can come up with in a script because the script doesn't have any way to know the angle an attachment is at.