[evlatests] WIDAR baselines, and time offsets

[George Moellenbrock]

>    1)  My error on the 'conjugation'.   Both real and imaginaries have
> their signs reversed to get an apparent 180 rotation.

NO!!  Changing the sign of both real and imaginary (globally) would
get you a negative image.

If the only problem were a reflected image (i.e., dynamic range ok), then
the answer is simply to change the sign of the phase for all visibilities.

I don't think this is the whole answer because the dynamic range is still 
bad.  I think a minority of baselines don't need the phase sign change (a 
sufficient minority that calibration sort-of works), and I think it is 
related to the antenna/baseline sorting....

Michael/Bryan: Is there a way to turn off the deliberate sorting done in 
the CBE?

-George

--