On Feb 27 2005, Bruce Lilly wrote:
On Thu February 24 2005 21:09, Laird Breyer wrote:
[...] users will
never have the kind of control over receipt of anonymous messages that
Bruce is positing.
I'm not positing any sort of control by individual users; rather I
am stating that organizational entities that have a need to accept
anonymous information -- and which have provisions for doing so now
by means other than Internet Messages -- will need to take appropriate
steps to ensure that they can do so. That may include careful
(administrative, rather than individual recipient) selection and
configuration of filters.
That's the kind of thing I'm saying can't be assumed in the presence
of a filter. I claim that the trend in spam filters is to be opaque to
users and administrators (super-users) alike, for both commercial reasons and
because of the increasing complexity under the hood (expanded in the
other message I sent recently).
Certainly, an organization which does its own mail handling can always
set up completely unfiltered addresses, and will have to painfully
sift through the junk. But if the addresses are filtered (possibly
upstream), then there can be no guarantees, since even the people who
build the filters don't understand fully how they operate.
For the purpose of discussing your draft, I doubt all this matters much.
Spam filters already operate outside of the "mail transport law" ;-),
and will continue to pick up clues they think are useful, regardless of
RFC semantics. Wondering whether they wil honour blah(_at_)[] is like
wondering whether they will honour a loophole. If spammers notice that
blah(_at_)[] gets through, then they'll all use those kinds of addresses and
filters will block them.
--
Laird Breyer.