The vertical and orthogonal \$L sub 1\$ linear approximation
problems for points are considered.  For the vertical case, a
optimal algorithm is given by using the prune-and-search technique
in a wise way.  The orthogonal case is related to the well-known
\$k\$-sets problem.