On Mit, 2011-11-23 at 14:55 -0300, Christian Grunfeld wrote:
[....]
> Flaws ?
> False positives....yes, ONLY the first time for each sender! just
> answer your good mails and they´ll become ham next time. Mails not
> answered (spam) remains as spam next and next and next !
1) That might look negligible for quite closed groups of people but if
you are on e.g. the Linux-kernel-mailinglist and some others near it,
it's quite different. After all, Cc: the LKML is mostly for the ML
archive and that Google can find it.
2) The next drawback is that some widely used MUAs do not show the email
address per default but only the comment near it (which should be the
real name but is technically even more comment than rest of the From:
field).
3) And I skim over the spam-folder once a week or so. Skimming spam (at
least) daily to actively look for ham is much more effort than just
making sure that an important (business or private) mail didn't got in
there (because of the sum all the usual little sins of the typical
Outlook-user).
4) And how does that - done in my MUA - integrate with SpamAssassin and
the like on the mail server?
evolution has nice buttons since ages. Does anyone know how "export" the
junk - aka spam - (and explicit not-junk - aka ham -) automatically to
the BayesDB for SA?
[....]
> Time to think in a new antispam paradigm !
The other problem is that probably many people don't like that everyone
thinks that they are spamming (even if it is true) unless the receiver
declares it ham.
I have no problem with that - actually it the way in real life with the
paper ads at the door and in the snailmailbox.
But I'm sure you are already trying and living by that witz email, no?
Bernd
--
Bernd Petrovitsch Email : be...@petrovitsch.priv.at
LUGA : http://www.luga.at
