[evlatests] software problem summary

[Ken Sowinski]

There are two problems we have learned about and fixed or
alleviated today.  Corrections or clarifications welcome.

1.  Wrong sideband
There were conditions under which the correlator system had
the wrong idea about the netsideband of the observation.  The
fundamental flaw was in the system controller since it was
commissioned last September; a more recent change to the
executor. probably in support of subarrays, manifested the
latent correlator problem.

Today the problem in the system controller was addressed at
the cost of reconfiguring the correlator when it is not
really necessary.  The bug in the executor was also addressed,
but there is a little evidence that it has merely changed its
nature.  More testing is necessary.


2.  Channel separation labels wrong
This problem appeared with the  with the idcaf that was started
on Spetember 13, and fixed today with the installation of the
updated executor and the September 18 version of idcaf.