Can anyone reccomend a linesort for use in conjugate gradient neural network error minimization methods? How about one which is easily parallelizable? -Thomas Edwards tedwards at cmsun.nrl.navy.mil ins_atge at jhuvms.BITNET