first do it one way and then if it is too slow after profling try another.
Like Knuth said "... premature optimisation is the root of all evil..."
so you are wasting time
trying to make something "more" effiecient when it is probably NEVER
going to be a bottle neck anyway.
And if a scripting element is your bottle neck you have much bigger
problems architecturally.
The scripting elements should be the non-preformance critical sections
of code anyway.
Get something working and then if it seems slow profile it and actually
see WHAT is slow.
