That's going to be the next step (just checking the Class C range), as most of the legitimate mail that I've seen that gets caught by this test comes from an IP that is in the same Class C as an MX entry.I agree with you completely. In fact, even with tiny clients, we often have the IMail mailbox server send directly, while the MX is on a different box. If the test included subnet-based weighting, it might be more useful; yet this would only apply to single-provider locations and not solve the problem of virtually or geographically distributed systems committing the crime of being well-architected!
Note that the test is still very useful, with the negative weight. Most spammer's mail will fail this test, so mail that doesn't can have a negative weight used (without hurting those servers that send their mail from a "distant" location from their MX records). The spammers whose mail doesn't fail the test are typically the mainsleaze spammers that can be caught based on their return address.
-Scott
---
[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.
