Preface -- I'm not a skilled C-language coder at all, just a hobbyist-artist. This is an absolute newbie question involving Cycles code, and I was referred here by staff at CGCookie.
Question: In the Cycles material node setup, the only displacement socket is the input socket in the final output node. I can see a use for intermediate nodes (separate from the material output node) to incorporate their own displacement "output" and "input" sockets, for things such as generating derived normals, the way various deform/generate modifiers appear to. [In other words, chaining several render-time displacements together as if they were modifiers in the mesh modifier stack, which is more accurate than linear adding of texture maps.]
If even a more skilled coder attempted such a thing -- would it be certain to introduce significant performance impacts; would the existing code structure even allow for it? Could it be done at the Python scripting level instead of hitting the C-level code?
This is my first time on any Stack Exchange site. If my question is poorly formed or if this is the incorrect venue for it, I can take advice on improving my approach for any future questions. Thank you to any and all who have read this far. [EDIT] (The intended output is the micropolygon dicer, not an image file.)