Jason Steiner wrote:
Why? If that single source is reliable and does a good job of implementing
the policies I choose for my mail servers (eg. "We don't want mail from
dialups here.") , what is the point of incurring the extra load of querying
more than one, especially if the second is going to be much more resource
intensive? (eg. A content filter behind a DNSBL.)
Indeed. Spam filtering is all about systems making informed choices for
their own needs. If system A thinks that a randomly generated "block
this" choice fits their needs, and system B thinks that 75 different
filters have to be unanimous before blocking, who are _we_ to judge
which is better? Not our machine. Not our users. None of our business.
Instead, we should be focusing on:
a) making sure that systems _can_ make informed choices by ensuring that
filtering methods tell the truth in a simple way of what they do
b) making suggestions on operational "properties" (such as consistency
to published standards, delisting mechanisms where appropriate) rather
than the blocking criteria themselves.
There is a DNSBL that blocks all of IPv4 space. There is another that
returns a randomly selected block/non-block indication on every query.
Are they bad/non-BCPish? No, because they're perfectly clear and honest
about what they do (including a comment about how dumb it would be to
actually _use_ them to filter email ;-).
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg