An implementation of the old adjacency method

Assume the same situation with the three adjacent stacks.  We'll determine 
all adjacencies at once.

c, c_l, c_r, P, A, K


A. As helper material, we can get an isolating interval for each section
   cell in the stack over c.  For each section cell we have an element
   of the form (s,J,K), where s is the signiture of the level 2 projection
   factors in the cell, and either:
   1) K = 0 and J is a logarithmic standard interval containing the cell's
      y coordinate and acting as an isolating interval for that value for
      each projection factor which is zero in the cell.
    
   - or - 

   2) J and K are standard logarithmic intervals contiaining no section 
      cells in the stack, and j2 = k1 = the y coordinate of the cell.

IPRRISD is messing up when faced with an input interval with an endpoint
which is a root of the input polynomial.