News for hackers. Stuff that matters.
Re:How Quaternions encode rotations: derivation and sample code by santaclaus on Wednesday September 07, 2016 @12:50PM (#12444944)
Isn't there some theorem stating that all 3 coordinate reps of rotations in R^3 have a singularity? But bumping up to 4 coordinates (or 9 if you like rotation mats) eliminates the singularity.

Edit: seems kind of like latitude and longitude. At latitude + or - 90, all longitudes map to the same point so that coordinate has no information.

Re:How Quaternions encode rotations: derivation and sample code by chombier on Wednesday September 07, 2016 @01:23PM (#12445271)
Yes, R^3 and the rotation group SO(3) are not homeomorphic (they have different topologies), meaning you can not cover SO(3) with a single chart. So any representation of rotations by 3 numbers is singular somewhere.