[daip] SNSMO bug

[Eric Greisen]

Many thanks - the problem has been resolved.  The program seemed to
think that it had to have both polarizations and so marked
(internally) a good one bad if the other was bad.  When you smooth
only amplitude, the MWF values for real and imag are then still bad
and so the point was skipped when rereading the table to flag it.  If
the phase was also smoothed then real and imag had non-blanked values
and so the point was examined.  When I drop the silly requirement for
both polarizations, then it works as it should with only ampl
smoothing set.

The MNJ tomorrow for 31DEC03 as well as 31DEC04 will have this fix.

Thanks,

Eric Greisen