Is there a de facto standard for applying translate, rotate, and scale transformations to a 2D texture on a cube face, for example? What do end-users expect?
I mean, should I scale, then rotate, then translate? Or use a different order?
Not to discover within ten years, I am doing something different from all other 3D graphics programs.
It is advised to first do scaling operations, then rotations and lastly translations when combining matrices otherwise they may (negatively) affect each other. For example, if you would first do a translation and then scale, the translation vector would also scale!
Consider too that scaling can be non-uniform, so if only one axis is scaled, you get less distortion if you scale before rotation, than doing it the other way around.