Corrigenda to "Undulant-Block Elimination and Integer-Preserving Matrix Inversion" Technical Report 418. The two corrections do appear in {\it Science of Computer Programming\/} which appeared in volume 33, number 1, January 1999. Vertical bars alert the reader to columns containing changes. Algorithm 2.4, second last line. | | Was: \item Recursively solve $U_{ne}\vec{z_n}=\vec{y_n}-E\vec{y_s}$. Should be: \item Recursively solve $U_{nw}\vec{z_n}=\vec{y_n}-E\vec{z_s}$. Figure 5, last alternative for backSubst. | | | Was: backSubst (Mtx (plusLU'nw,_,w,plusLU'se)) (Vec y_n y_s) (Vec omega_n omega_s) = (Vec z_n z_s) where z_n = backSubst plusLU'n (y_n - e #*| y_s) omega_s z_s = backSubst plusLU's y_s omega_n Should be: backSubst (Mtx (plusLU'nw,e,_,plusLU'se)) (Vec y_n y_s) (Vec omega_n omega_s) = (Vec z_n z_s) where z_n = backSubst plusLU'n (y_n - e #*| z_s) omega_s z_s = backSubst plusLU's y_s omega_n