Original Post
I want clarify that I understand things correctly. am i correct in saying that opengl does left-multiplication using column-major matrices when using the standard transform calls, glRotate, glTranslate etc. so if I wanted to scale(S) then rotate(R) then translate(T) a point P (row vector) (where S,R,T are row major) I would usually call
glTranslatef
glRotatef
glScalef
draw P
with the appropriate parameters. And taking R,S,T to be row-major, opengl's calculation looks something like T'R'S'P' which is equivalent to PSRT where ' denotes the transpose (swaps to column major). So if the current stack is given by W such that its effect on a point P is W.P' any new operation (row-major R) added to the stack, produces: WR'P' Am I correct in my thoughts? rewolfer edit: when i say equivalent, i mean value-wise - not orientation. the one produces a column vector, i know.