On Sun, Dec 15, 2019 at 06:04:59PM -0500, Keith Moore wrote:
> I do not find that this is true in general, and IMO it is not a valid
> working assumption. There's a tremendous amount of Bad Conventional
> Practice out there, especially in both email handling and security,
> that has come to be accepted as if it were wisdom.
Yes, there is much cargo-culting of questionable practice. But there
are also in some cases specifications that (no longer) match operational
realities. Which is not to suggest that support for address literals is
necessarily such an example, it rather depends on metrics I don't have,
and which may even be time and place specific.
Perhaps there was once a particular stream of abuse directed at ietf.org
that the rule in question stopped. It may still be effective, then and
could now be merely a low-volume source of false-positives. Or it could
still be, perhaps intermittently, an effective way to distinguish spam
from ham.
If I were the operator of the MTA in question, I'd consider dropping the
rule after checking recent logs for signs of its efficacy.
--
Viktor.
