Ricardo Kleemann wrote: > > I'm actually trying to use dspam in addition to SA, hoping maybe that it > picks up the spams that SA isn't.
I don't use dspam, but it occurs to me that if you run dspam in combination with spamassassin, your users won't be getting enough actual spam to be able to give dspam meaningful training. IIRC, in order for it to learn to classify messages, it needs substantial input of both good and bad messages. If spamassassin ultimately blocks most of the spam, then users can't feed that information back into dspam to train it. I'd like to know if I'm wrong, but I expect it will be of very little use. I've been blacklisting based on the spamhaus Zen list (in esmtpd) for most of my customers, universally increasing the score for the SURBL tests and spamhaus tests in spamassassin, and using Razor, Pyzor and DCC with very good results. Spamassassin does all of the tests, and is called from pythonfilter. I add these lines to local.cf: bayes_path /var/lib/spamassassin/bayes/bayes bayes_file_mode 0666 auto_whitelist_path /var/lib/spamassassin/awl/auto-whitelist auto_whitelist_file_mode 0666 pyzor_options --homedir /etc/pyzor razor_config /etc/razor/razor-agent.conf score URIBL_AB_SURBL 2 score URIBL_JP_SURBL 3 score URIBL_OB_SURBL 2 score URIBL_PH_SURBL 2 score URIBL_SC_SURBL 3 score URIBL_WS_SURBL 2 score RCVD_IN_SBL 5 score RCVD_IN_XBL 6 score RCVD_IN_PBL 2 ------------------------------------------------------------------------------ Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA -OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise -Strategies to boost innovation and cut costs with open source participation -Receive a $600 discount off the registration fee with the source code: SFAD http://p.sf.net/sfu/XcvMzF8H _______________________________________________ courier-users mailing list courier-users@lists.sourceforge.net Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
