About the Rotation in the Anisotropy node… shouldn’t be the environment light to determine the direction in which the Anisotropy is stretched over the object surface?
You can certainly rotate your light in 3D space to emulate the effect. But that should really only be applicable to a rectangular area light as all other light sources are singular points in space and rotating will have no effect.
I'm not sure if including the rotation parameter in the anisotropic shader is scientifically logical or just convenient for we artists.
I wasn't talking about rotating the light, but whether the lights of my scene should determine the Anisotropy's stretching direction. I thought the effect was related to how the light hits the object, but it is actually an intrinsic quality of the material, am I right?
It's both. Primarily I would light my scene to look good with or without my anisotropic material(s). Once that's good, I'd add the anisotropic material(s) and alter the rotation value in the shader rather than the lights. There's more consequence to the overall render if you adjust the lights to accommodate the desired anisotropy; less risky to simply adjust the material rotation setting.