mark lewis wrote in a message to William McBrine:

 ml> the easiest option is to use MSGID and simply use a serial number
 ml> that starts at x and increments x+1 for each message posted... the
 ml> only problem with this method is determining what digit to start x
 ml> at because it is possible that a system may need to be reinited and
 ml> thus start with a new counter...

The gateway program I run here (GIGO) uses a file for this purpose:

   sequence.gig  8  02-Jun-03 20:53

LIST,  in hex mode,  shows me this:

000000  6D 11 00 00 C8 03 00 00

 ml> of course, we also have to look to the buggy and flawed software
 ml> that regurgitates messages with modified headers and control
 ml> lines... those that strip, replace or add MSGID will mess up thie
 ml> method but many would rather read a true duplicate a second or
 ml> third time than risk loosing a message and having it never get read
 ml> at all...

I used to have a message area set aside for dupes to be tossed into,  but
instead have them deleted these days.  Every once in a while I'll see some
in the logs,  and wonder...

And speaking of buggy and flawed software,  I can still remember occasions
that go by every so often where some system out there will dump *ALL* of
their messages into a whole mess of echos.  The first clue is that they're
(usually) all dated a ways back,  maybe as much as a couple of months,  but
not always.  The other good clue is that most of them show the original
from/to headers and origin line,  but if you look at the PATH the first
system there doesn't match the origin line,  not at all.  And the PATH
lines are all the same...

I'm surprised that somebody hasn't come up with a fix for this one by now.

---