a # 1 60 R ' r#u drounds v;Lct;Lio;e;n                                  '
a # a,[1]  ' e#1|v#v%u+Lct#Lio#0                                        '
a # a,[1]  ' n#(D0.5++/v)_+/Dv                                          '
a # a,[1]  ' r#uX(Dv)+n>AWe                                             '
a # a,[1]  ' $0                                                         '
a # a,[1]  '"drounds - distributive rounding of a vector to arbitrary   '
a # a,[1]  '"scalar unit                                                '
a # a,[1]  '"from ''the apl handbook of techniques'' copyright 1978, ibm. '
a # a,[1]  '"-rounding the elements of a vector before summation may    '
a # a,[1]  '" cause an error in the sum.  round-off errors do not       '
a # a,[1]  '" necessarily compensate.  it would be good practice to     '
a # a,[1]  '" carry maximum precision until the final summation, then   '
a # a,[1]  '" rounding the sum.                                         '
a # a,[1]  '"-when this is not possible, we would still want the rounded'
a # a,[1]  '" sum to equal the sum of the rounded elements.  see Grounds'
Lfx a