Is there an inconsistency between this article and the actual implementation of Turn in PyMOL? Reading this article gives the impression that the camera rotates about its axes at its origin in camera space. If this were true, an issued command such as "cmd.turn('y',180)" should actually move an object placed before the camera (perhaps centered with its position in +z, for instance) to a position *behind* the camera (-z). In practice, that is not what I observe. Instead the object of interest appears to rotate about its own central axes.
I have posted here:  An example of view matrix alteration produced by Turn, in this case a 180 degree rotation about the y-axis.
To me it seems that what it actually being effected is the alteration of the relationship between model and camera spaces rather than applying a rotation matrix to the object coordinates in camera space.
My recollection of the involved mathematics is a bit vague, but these are relatively intuitive concepts of notable distinction. I'll brush up on the math, but I think the problem lies with a misunderstanding of PyMOL's workings or a lingual inaccuracy of the article. I am relatively new to PyMOL and only recently have started peering inside its source code. Perhaps my understanding of its workings is flawed? If so I hope some discussion can help inform how to improve the PyMOLWiki.
--SavinaRoja 20:44, 7 December 2010 (UTC)
I think you are right, turn does more than rotating the camera. Maybe it could be phrased like this: The camera is rotated about its axis and afterwards, its position is adjusted such that the origin of rotation stays at the same position. Or (maybe better?): The view of all objects is rotated about an axis that cuts the origin of rotation and is parallel to the specified axis. Maybe this is what "centered at the origin" is supposed to mean.
Or maybe just "The view of all objects is rotated about the specified axis". This is just what most people need it for and maybe they are not interested in camera space, object space, and model coordinates...
I've been thinking about these things so much because I needed changed model coordinates but still I find it complicated...
--Karo 09:48, 9 December 2010 (UTC)