I am very unhappy with the unflexibility of user defined functions. Recursion is
not allowed, loops can not be implemented, the only way to branch is the select
function, which is very unhandy and I am not sure whether POV-Ray still
evaluates all branches and after that selects.
Furthermore I would like to be able to write to files from inside a function
during render time, not just while parsing.
(Please correct me if anything I said is wrong and actually possible)
The best way I see for a maximum of flexibility while maintaining (or even
enhancing) render performance would be the ability of including external
functions, defined in DLL libraries (or the Mac/Linux equivalents).
One could even think of a new type of surface where POV-Ray does not try to find
the isosurface of a 3D-function itself but hands the start point and the
direction vector of a ray to a function and expects the intersection point.
In my opinion this would be a great feature for POV-Ray, enhancing its (already
amazing) versatility a lot. And I think the implementation can't be that
My problem is, that I have neither patience nor time to become familiar with the
POV-Ray source and implement this feature myself.
I hope that someone who already knows the source likes my idea and implements
Post a reply to this message