Mailinglist Archive: opensuse (3232 mails)
| < Previous | Next > |
Re: [SLE] Whats with thi8s spamassin
- From: "jdow" <jdow@xxxxxxxxxxxxx>
- Date: Wed, 2 Aug 2006 07:39:38 +0000 (UTC)
- Message-id: <049701c6b606$c71c7db0$0225a8c0@Wednesday>
From: "Peter Nikolic" <p.nikolic1@xxxxxxxxxxxxxx>
It seems that SUSE is running SpamAssassin and does not have BAYES
very well trained. You also MIGHT be triggering other filter elements
if you the email is going direct from your machine to SUSE rather
than going through btinternet smtp servers. There are a lot of BLs
that list blocks of IP addresses that are assigned dynamically. These
get modest spam scores as a general rule.
It us a hassle. The cure for it comes if they use different BAYES
setups for each list and carefully train them to recognize typical
list traffic as ham and other stuff as spam. Since the list is mostly
ham it is easy to get BAYES rather lopsided if they try to train
BAYES dynamically. (I hand feed my own BAYES filter here, choice
morsels only.) MAYBE it will straighten out over time. But it is
rather frustrating to the victims of badly setup filtering.
{O.O} Sympathies. Joanne
Hi .
I just posted to the list for the first time in a couple of weeks and now see my post marked up as SPAM by list WHY ..?....
For those that may not have bothered to look my email address is @btinternet.com NOT yahoo.com or .uk it would be nice to see this unesercerry (SP) hassle sorted PDQ ..(list admin are you awake ?..)
It seems that SUSE is running SpamAssassin and does not have BAYES
very well trained. You also MIGHT be triggering other filter elements
if you the email is going direct from your machine to SUSE rather
than going through btinternet smtp servers. There are a lot of BLs
that list blocks of IP addresses that are assigned dynamically. These
get modest spam scores as a general rule.
It us a hassle. The cure for it comes if they use different BAYES
setups for each list and carefully train them to recognize typical
list traffic as ham and other stuff as spam. Since the list is mostly
ham it is easy to get BAYES rather lopsided if they try to train
BAYES dynamically. (I hand feed my own BAYES filter here, choice
morsels only.) MAYBE it will straighten out over time. But it is
rather frustrating to the victims of badly setup filtering.
{O.O} Sympathies. Joanne
| < Previous | Next > |