We use an M based mail program called VA Mailman to distribute software
updates. Occasionally, a message that has already been sent will be purged
from the target system and it will have to be resent. Unfortunately, the
other system will have recorded that it has already sent message #nnnnn to
domain wherever.va.gov, and so it will not resend the message. Before this
problem was addressed, it was necessary to manually copy the message and
then send the copy. This situation arises because we store all messages in a
single database and each recipient on the local system simply stores a
reference to the message, and the message ID is the same for everyone. Now,
I'm not saying this necessrily the right way to do things, but it is an
example of how things can break when the policy of not resending the "same"
message is enforced a little too aggressively (or, perhaps more to the
point, the definition of sameness is too exacting).
===
Gregory Woodhouse <gregory(_dot_)woodhouse(_at_)med(_dot_)va(_dot_)gov>
Financial Product Line
+1 415 744 6362
"Computer science is no more about computers than astronomy is
about telescopes." -- E.W. Dijkstra
-----Original Message-----
From: Gregory Sereda [mailto:gsereda(_at_)maillennium(_dot_)att(_dot_)com]
Sent: Wednesday, May 03, 2000 9:23 AM
To: Randall Gellens; Randall Gellens; Tim Showalter;
ietf-mta-filters(_at_)imc(_dot_)org
Subject: Re: last chance before last call on sieve and vacation
At 08:54 AM 5/3/00 -0700, Randall Gellens wrote:
... and never send a message to someone twice unless the :days
specified in the vacation message has been exceeded or the text to
be sent has changed.
The wording "a message" implies "the same message". Since the sentance
is getting a little long, the editor may consider this wording...
... and never send the same message to someone twice unless the :days
specified in the vacation message has been exceeded.
Greg Sereda
AT&T Labs