On Fri, 25 Apr 2003 jwsacksteder@??? wrote:
> It would be very useful to have a callout verification option to use
> VRFY instead of RCPT when verification is called for.
We had this discussion last week. Here is what I posted on April 21:
----------------------------------------------------------------------
I do not like this idea, for this reason:
. The point of sender verification is to answer the question "can I
deliver a bounce message to this address?" Successful routing gives some
of the answer; callback helps with additional information.
. The point of recipient verification is to answer the question "can I
delivery this message to this address?" Again, successful routing gives
some of the answer, and callforward helps with additional information.
The delivery commands are MAIL followed by RCPT. That is why Exim uses
those for callouts. They give some assurance that delivery is likely to
be successful. There is no such assurance for VRFY, a command that many
sysadmins lock out. Other MTAs always answer "yes" to all VRFYs, which
makes it pretty useless. Indeed, it is now recommended that the "don't
support" response to VRFY is 252 rather than 5xx because of such
behaviour.
----------------------------------------------------------------------
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
