use the same coordinate system please

Lindens,

please ... if you use a z up system for translations don't use a y up system for rotations for chris sake. (v 1.6.3)
And please let the primitives be rotated arbitrarlzy ... not if I rotate the around x axis that the z axis rotation gets changed (this happens when the y axis is rotated 90 degrees).

I think you might be mistaken in your troubles. I just tried rotations in 1.6.3 and do not see what you are reporting so I'll try to see if I am reading you right.

The Z axis is the one pointing up out of the ground, you know that. Rotations around the Z axis will rotate a prim around the Z axis in the X-Y plane. However, spheres are created by default with their local (prim) +Z axis pointing where +X usually does for cubes (i.e. 90 degree rotation around Y). I think this is done for the reason that a sphere textures more as one might expect when rotated 90 degrees. This also applies to the roundish prims: half spheres, rings, torii.

In 1.6 there was a change in the way coordinate axes are shown, there is now a pulldown in the edit box which allows "World" (the "Cartisian" space) "Local" (the prim or link set axes) and "Reference" (which I can't figure how it is different from World). If you are set to "World" ruler type and rez a sphere and edit it, the arrows will not show you that the sphere is rotated and so then your manipulations may well be confusing. Switching to "Local" ruler will show you how the prim axes are oriented and may help you with the problem you are having.

Rotations in the edit box are weird because rotations are stored internally as four element quaternions but are displayed in the more human friendly X/Y/Z degree rotations. Math is not my strong suit, but I think that there are many (infinite?) quaternion representations for a given X/Y/Z "Euler" rotation. It is the conversion between quaternions and Euler rotations that causes the edit box to behave "oddly", but this has been the case since the start of SL. As noted on the wiki page referenced above"Quaternions are the things that scare all manner of mice and men. They are the things that go bump in the night. They are the reason your math teacher gave you an F. They are all that you have come to fear, and more. Quaternions are your worst nightmare". Forewarned is four armed, in this case.

This is a problem that's been introduced in 1.6 - the "default" for the ruler is now "World." This can be slightly confusing to the layman - but, yes - setting the option to "Local" helps. Occasionally it may be a little buggy, though... if this happens, select the prim again.


Correct. Again, for the layman, Eulers are usually the easiest to understand for the layman.


This problem is known as Gimbal Lock. It happens in-world often if you try to rotate along the Y axis.


This quote is from here: http://www.cprogramming.com/tutorial/3d/quaternions.html

Are you sure?

I though gimball lock was what heppens with a real world turret...

As you rotate and elevate the barrel, the barrel is restricted from tilting/rotating to either side. Quaternians are usefull in a 3d environment because they don't display this behavior.

Objects therefore move more fluidly in the environment because of this. Makes it a stone b*tch to make a realistic behaving turret though as the barrel flops over in very annoying but at least predictable ways when you want to perform both a rotation and elevation change together...

Argh...

Racer P.

Rez a box. Rotate arbitrarily. Use menu Tools>>Use selection for grid

Rez another box, choose Reference as Ruler mode.

The reference axes are identical to the local axes of the first box (the reference box). In other words, when you rotate the second box to <0,0,0> on reference axes it will have the same rotation as the reference box. Movement and scaling also uses the reference axes (the dimensions of the ref box are steps on the grid).

Sadly, this only applies for direct manipulation using the handles/grid. The number input boxex (also rotation numbers display) still show the world numbers

Actually, this is the way it should be.

If you picture a line through the Z axis of a cube, it is going up and down. If you rotate on the Z axis, it rotates around that line. If you rotate on the Y axis, it rotates AROUND the Y axis.

This is how it should be.