Uprgraded to critical, this one is pretty well broken. Havok4 should NOT be released before this is function is fixed.

Thought it could fit into meta issues such as "Scripted Avatar Movement"

I'm actually starting to get the feeling there's a handful of underlying physics issues all acting up at once. One of them being "Objects at rest, stay at rest"

http://youtube.com/profile_videos?user=n1nj4Ryozu

It seems that
Havok1: Movement is dampened over time, so velocity is a bellcurve
Havok4: Velocity is constant up to the point you reach the target, at which point you kind of overshoot, and are sent back the other way.

If a catchall or meta jira item is created, the pieces should be made as sub-tasks.

Of special note is the behavior of the plain object in Havok4 seeming to.. jump around strangely... I'm not sure if it failed to make it to the target, and required some kind of interaction (My avatar influencing it) or if it was the client failing to accurately show the position until some kind of update was forced.

Here's the relevant code used in each object. I hard coded the positional data so that if I didn't actually make it to the exact position, I wouldn't drift between clicks.

/////////////////
integer On;

default
{
state_entry()

touch_end(integer total_number)
{
if (!On)
{ llSetColor(<0.0, 1.0, 0.0>, ALL_SIDES); llMoveToTarget(<175.0, 175.0, 50.0>, 0.05); } else { llSetColor(<1.0, 0.0, 0.0>, ALL_SIDES); llMoveToTarget(<175.0, 175.0, 55.0>, 0.05); }
On = !On;
}
}

//This was placed in two spheres set to a scale of <0.25, 0.25, 0.25>
//One sphere has it's X position in each llMoveToTarget incrased by 1.0
//to avoid collisions with the avatar.
//Each click moves the avatar up or down. Make sure the Z positions are above
//ground and 5.0 meters apart.
////////////////////////////////

CrystalShard Foo, the sagging effect of llMoveToTarget(llGetPos(), 1) is expected. The value of 1 at the end means that the motion will be critically damped with a "timescale" of about 1 second --> the object's offset from its target is approximately 1/e of what it was one second ago --> when it is rather close to its target the effect is weak. Unless you also give the object buoyancy it will fight against gravity and will therefore sag a bit from its target position when the "timescale" of the effect is long (1 second). Alternatively, you could give it a smaller timescale (0.1 second for example) and it will try to get there very fast and sag less against gravity.