Magnus Sundberg wrote:
How do you know that the content of the deleted mail was spam?
Sorry if I didn't express myself good enough, I'll try again:
Spam is delivered to the user's SPAM folder. Good mail is delivered to
the users INBOX. That's how it works.
Now dspam can make a mistake, especially in the beginning of the
training. There two cases where dspam can be wrong:
1) either it thinks a mail is good, whereas it was actually spam or
2) it thinks a mail is spam, but it's actually good.
The first case is basic training, telling dspam what is spam. The second
one is correcting it's mistakes.
1) If dspam missed a spam mail and delivered it to the default INBOX
folder, the user moves the spam mail to the SPAM folder. Trigger goes
off and dspam relearns that this mail was actually spam.
2) If dspam got a false positive and delivered a good mail to the SPAM
folder, the user moves the false positive from the SPAM folder to
wherever he wants it (INBOX, My_documents, Family, whatever custom
folder). Trigger goes off and dspam relearns that this mail is actually
a good one.
The user moves something _to_ and _from_ the SPAM folder only in these
two cases. Deleting is still deleting, we don't care about that. Only
moving from and to the SPAM folder matters.
I hope I cleared that up now,
Alex
