All right,  please don't  send more  evidence. In  addition to  the files
duplicated on  the SEARN-SEGATE connection,  there was a  second problem.
When I saw the duplicated files in the JNET logs I assumed of course that
this was  the cause of  our problem, when in  fact it only  accounted for
some of the duplicates. The current local timestamp is:
 
                     Thu, 13 Apr 1995 02:44:26 +0200
 
If you  receive duplicates  with a  "Date:" field  posterior to  that, it
means we have  a third problem or  the fix for the  second problem didn't
work. The second problem has been around for about a year but only occurs
during  unusual,  timing-critical  circumstances.  This is  the  kind  of
problem that sits in a corner for months without ever touching a message,
and then  suddenly springs up as  the right conditions are  met. It gives
programmers nightmares because  everything seems to work  fine for months
and the next thing you know, your system is dead :-)
 
  Eric