[daip] new data weights from FITLD

[Eric Greisen]

I just checked the origin of these:

The flagging is done on tape weights.

The SCLVIS parameter was based on a scaling found under the name
VIS_SCAL in the UT (vis data) table.  It is after CORVER 4.17
multiplied by the integration time in seconds. From the code in one
place, it states

C      SCLVIS   R      The visibility scaling factor. If a visibility
C                      measured by the VLBA has weight w then the
C                      actual fraction of the integration period over
C                      which the data were acculated was w/SCLVIS and
C                      0.0 <= |w| / SCLVIS <= 1.0.

This SCLVISnew = SCLVISold * int_time

 and

 w = w_tape * SCLCISnew.

New code can divide the weight by the value in the random parameter
pointed at by ILOCIT.

Eric Greisen