Things that I've done for the past week:

1. Prepare the lecture slides for the STA.
   I've done the first part on static timing analysis.
   I'll work on the statistical timing part early this week.

2. Theoretical proof of the theorem for the ICCAD submission
   When I was giving the detailed proof for the first
   two theorems, I found that there was an argument that
   was not correctly used. Therefore, the theoroem One and Two 
   now become not quite right for arbitrary distributions. 
   
   My intuition is that to make such a property to hold I shouldn't
   make it too general (i.e., work for any PDFs). Now I'm thinking
   to refine those two theorems to make it work. 

   Fortunately, Theorem 3 and 4 still hold, which are the main
   foundataions for my implementation. Therefore, the overall
   conclusions in the paper are still valid.