"Jon Kyme" <jrk(_at_)merseymail(_dot_)com> wrote:
Yes, this has been said before. However, practically,
is_spammer(reputation_service, cut_off_metric, sender)
is exactly equivalent to
is_spammer(whitelist, sender) for any cut_off_metric.
Which is why you don't ask "is_spammer()". This was discussed on
the SMTP-VERIFY sub-list a few months back. Binary decisions are a
step above useless, but not a large step above useless.
Instead, the question should be "get_metric(service, sender)". This
permits the recipient to make it's own decisions. It also prevents
the reputation service from making your decisions for you.
For a real-world example of this working, see the FIDO political
wars in the 1980's. The node list was distributed to all
participants, so when a group tried to take over FIDO, they couldn't,
because everyone had the information for themselves. If FIDO had been
designed with a centralized list, then the administator of that list
could shut it down at any time.
But this is off-topic for MARID.
Alan DeKok.