Step 1: remove all the lighting stuff (I don't know why you based it on the lighting shaders in the first place), this will save atleast 20 slots
Step 2: pack all varying variables into vec4 types. each varying will use up 4 slots regardless of its type, that is, 4 * vec3 will use up 16 slots but 3 * vec4 will only use up 12 slots for the same total number of components (12).
Et voila, now you should be well under 64 slots and it should work on most cards.
I'll do that, thanks. As for the overflow, I had only planned to add extra 'fake' lighting contrast to scenes initially to counter the scaling on sprite BG's in FF7, when you're running the game at very high res, the models clearly stand out from the 2d background, and this sorts it.
Of course, when you start someting you mean to do very little on, you end up going a bit overboard and adding too much lol.
Edit: Great work on the driver btw, very impressive coding. I assume you hear this question too oftan but, do you have any plans on releasing the source?, I'd love to tinker around with it.
