I think this is the “virtual trackball” from this SIGGRAPH 1988 paper http://www.math.utah.edu/~palais/Links/Chen_Trackball_3-D_rotation.pdf http://www.math.uta

I think this is the “virtual trackball” from this SIGGRAPH 1988 paper http://www.math.utah.edu/~palais/Links/Chen_Trackball_3-D_rotation.pdf http://www.math.utah.edu/~palais/Links/Chen_Trackball_3-D_ro... which I think was also the method described in Graphics Gems circa 1990. A mouse is in general a horrible input device for inputting arbitrary 3d rotations, because it doesn’t have enough degrees of freedom. By far the best tool is a 3-DOF trackball set up so that arbitrary rotations of the ball are directly converted into rotations in the computer. Next best is some other physical device with gyro/accelerometer so that (slightly less accurate but still) direct manipulation can be input. Next best after that is something with multitouch or perhaps one of those 6 DOF “space mouse” devices.