Notice that in the example the entries in the matrix L are the negatives of the multipliers used in the row operations. This is no accident, as we will soon see. Thus if the entries of both L and U are chosen to be integers, the multipliers and all entries of A will be integers, so then the arithmetic will be in integers as well, which is what we wanted. The method is based on the L-U factorization, which is now creeping into standard linear algebra textbooks. A good general reference is Strang [10] or Tucker [11].