[Mnj] CVEL patch

[Eric Greisen]

The 31DEC09 amd 31DEC10 code of CVEL after 20 September 2009 lost the 
logical variable that told a bottom level routine to compute velocities 
for the VLBA based on the center of the Earth rather than the antenna 
position.  Thus CVEL has been doing the wrong thing for VLBA data since 
that time.  The code in 31DEC10 has been corrected and will be available 
by MNJ tomorrow.  The code in 31DEC09 has already been corrected and may 
be downloaded as a patch or by MNJ now.

This does bring up a major issue in CVEL.  That task assumes that the 
fringe stopping in the correlator shifts the frequency to that 
appropriate for one of the two antennas.  CVEL uses the lower numbered 
antenna for all arrays except EVN and, for both the old and new VLBA 
correlators, it ignores the antenna location.  I am told that some 
non-VLBA correlators have changed to use the Earth center also but CVEL 
does not now support that.  Furthermore, there is a danger that antennas 
will be renumbered so that the lower number becomes the higher and vice 
versa which makes a mockery of the above algorithm.  Any further 
information you can provide about what correlators actually do should be 
helpful.

Eric Greisen