Scott,
I noted the following ROUTING hit on a message that went from public IP space to private IP space and back to public IP space. I'm not sure how often this happens, but maybe it would be a good idea to disable the test when it has private IP space within the path? Here are the headers:
Received: from wip-ectls-mx1.wipro.com [203.101.113.37] by mx1.mailpure.com with ESMTP
(SMTPD32-8.05) id A28A48F80238; Mon, 20 Sep 2004 15:42:02 -0400
Received: from wip-ectls-mx1.wipro.com (localhost.localdomain [127.0.0.1])
by localhost (Postfix) with ESMTP id 155BE220053
for <[EMAIL PROTECTED]>; Tue, 21 Sep 2004 01:11:55 +0530 (IST)
Received: from blr-ec-bh2.wipro.com (blr-ec-bh2.wipro.com [10.200.50.92])
by wip-ectls-mx1.wipro.com (Postfix) with ESMTP id F1305220004
for <[EMAIL PROTECTED]>; Tue, 21 Sep 2004 01:11:54 +0530 (IST)
Received: from usa-bos-msg.wipro.com ([205.163.139.2]) by blr-ec-bh2.wipro.com with Microsoft
Basing the test on the seperate IP database that you distribute might have an even better effect on the accuracy of this test.
Related to this but involving other technical tests, it seems like this test and others like BADHEADERS with it's very bad false positive problem with Outlook clients has become somewhat forgotten and much more unreliable over time and it might be a good idea to straighten the tests to limit false positives. The MailFrom test has even become somewhat unreliable due to the introduction of sub-domains and a test for a sub-domain IMO could and should be scored differently than from the base domain.
Markus' stats show some very concerning patterns with these tests: http://www2.spamchk.com/public.html . Some of this I think could be based on his location and the predominance of non-English languages and not-as-popular mailers/clients (China has big issues with this for instance from my own experience), but it definitely suggests that the conditions have changed significantly causing many of these tests to become less accurate and that could use some attention in the future in order to shore up, pending the launch of a new release of course.
Thanks,
Matt
-- ===================================================== MailPure custom filters for Declude JunkMail Pro. http://www.mailpure.com/software/ =====================================================
--- [This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)]
--- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type "unsubscribe Declude.JunkMail". The archives can be found at http://www.mail-archive.com.
