Posted by: Anonymous Coward
on October 29, 2003 11:15 PM
In the comments section of your previous article, you dismissed script-based (Perl, Python, etc.) solutions mostly because performance issues against compiled C/C++ code.
However, it seems too much hassle to manage the dictionary for the junk filter with Annoyance Filter. Almost every other anti-SPAM GPL solution out there offers some way to analyze incoming e-mails by itself and in some cases, like POPFile, offers a nice and intuitive interface so the user can deal with it.
Even the procmail rules for SPAMAssassin (which can work at the MTA or/and the MUA level) don´t seems so hard (perhaps "involved" would be a better word here) to manage in comparision with this.
While I´ll concede that performance issues should be taken into account, today´s machines are powerful enough to run those script-based anti-SPAM solutions without affect the responsiveness of the others applications and the OS overall.
I do think that such software should not be too much intrusive. Better yet, the user shouldn´t be too involved with the process of managing the classification rules. These things counts a lot in my book when I have to manage the rules on a timely basis.
Too much hassle!
Posted by: Anonymous Coward on October 29, 2003 11:15 PMHowever, it seems too much hassle to manage the dictionary for the junk filter with Annoyance Filter. Almost every other anti-SPAM GPL solution out there offers some way to analyze incoming e-mails by itself and in some cases, like POPFile, offers a nice and intuitive interface so the user can deal with it.
Even the procmail rules for SPAMAssassin (which can work at the MTA or/and the MUA level) don´t seems so hard (perhaps "involved" would be a better word here) to manage in comparision with this.
While I´ll concede that performance issues should be taken into account, today´s machines are powerful enough to run those script-based anti-SPAM solutions without affect the responsiveness of the others applications and the OS overall.
I do think that such software should not be too much intrusive. Better yet, the user shouldn´t be too involved with the process of managing the classification rules. These things counts a lot in my book when I have to manage the rules on a timely basis.
#