User blog:Koteitan/Concestor Approach for Searching bad part of Bashicu Matrix Systems

Though I am thinking of change decision of the bad root (=D-value of Buddy3's) to fix Bashcu Matrix System version BM1 as like Bubby3's I found another approach.
 * 1) Write all rows in hydra notation.
 * 2) Mark all ancestors of the rightmost nodes of each rows.
 * 3) The rightmost node which is ancestors in the all rows (=concestor) is decided as the bad root (=D).

With this approach, (0,0)(1,1)(2,2)(3,3)(4,1)(3,2)(4,2)(5,2) can have the bad root (0,0) instead of (1,1) which is not the ancestor of the 2 of (5,2) in the second row (B in the figure below). I think Buddy3's approach can have mistake of the selection of the ancestors by erasing branch for the upper rows and making the illegal connection for the lower row.

With this bad-root-searching method, each bad part in each rows is copied on a single branch of the each hydras. I think this change helps the proof of termination.

Though I post my idea and implement the C program in my github before, But I have the bug in the implementation... I will fix it soon.