The "pivot" in cracktransform() and matrix?

[aureolekin]

       Hi guys, I'm Aureole. I've been trying to wrap my head around Matrix and transformation stuffs lately, but there's still something intriguing to me that I can't figure out, especially pivot and vector  cracktransform(int trs, int xyz, int c, vector pivot, matrix xform) function! For example:

 

1. Where is the "pivot" information stored in a 4*4 matrix ?

2. Apparently typing different "pivot"  in the cracktransform function yields different results. Then is there a "correct" pivot to use in the cracktransfrom() function to get a meaningful result?

3. I know that the @pivot intrinsic attribute will be created when a packed prim is created, and this attribute might be the "correct" pivot information to use in the cracktransform function. However, for regular non-packed geometry, is {0,0,0} always the "correct" pivot information to use even if the geometry is not centered at origin?

　

      I don't know if I've expressed myself clearly, pls let me know if you don't get me Many thanks!