> Then why not simply use an email confirmation like is used
> for verifying
> subscription and then cache the address for n days (where n
> is 30 or so)
> if the response comes back positive? Or, even better, only
> require email
> confirmation for those messages which get scored as positive for spam.
>
> Spam only works because its one-to-many. Forcing a
> confirmation requires
> many-to-one and will collapse a spam server (or be ignored,
> either way you
> win). In addition, adding a two-way element avoids all the
> ethical issues
> of blacklists. If someone really wants to post to the list
> that badly and
> they score positive, they'll send the email verification.
Maybe spammers will adapt to that way of confirmation some days.
Then it would be further interesting to have a human to review
any such non subscriber contributions. The amount should be
minor when the few persons that cannot always post from the
subscription address are set on a "good list".
-Alex.
PS: you will never be able to driver a mailing list without supervision
because its a "nice and needed multiplier" for spammers.
