On Mon, 22 Sep 2003, Larry Smith wrote:
Have followed the thread and am aware of what has been said. Point still
remains, your comment is that in "neither" case will the message get
delivered - and my comment was "not totally true". I am not referring to the
MUA doing the check, I am referring to it being done at the MTA upon hand-off
from the MUA. Pre-Verisign "many" mail servers (mine included) would
"reject" a bogus or non-existant address _before_ receipt of the message from
the MUA therefore leaving the message on the client/customer machine (at the
MUA).
Sure, but when the MTA rejects something in this case, a bounce is
returned to the user. As I said, in all cases a bounce should be returned.
In this case, as bounce is still returned. The only (trivial) difference
is that the MTA must make a connection to the verisign server to get its
rejection.
All that has been changed (at most), is the error message in the bounce
message.
However, the configuration you describe has been shown to be less than
ideal. High speed systems first queue all the messages, then filter spam,
then route those that are not spam. Attempting to reject messages "early"
results in poorer performance. So I think we need not be too concerned
with the nominal effect on such systems.
--Dean