||In the article's "Another easy solution in <o logn o>" of LCA, it uses dynamic programming, with the following recursion:
When j = 0, then: P[i] [j] = T[i];
When j > 0, then: P[i] [j] = P[ P[i] [j-1] ] [j-1]
Since P[i] [j] is the (2^j)-th ancestor of i, when j > 0, I think P[i][j] should be:
P[i] [j] = P[ P[i] [j-1] ] 
Please correct me if I am wrong. Thank you.</o>
Updated on 10/06/2015:
Never mind, the original version is correct.