Stavros Macrakis wrote: > That is not the bad case that was reported (though it is another bad case). Also, should we round the printed result? fpprec:16; fpprintprec:4; 1.56789b0 -> 1.567b0 Ray