[EMAIL PROTECTED] (Justin Mason) writes: > Could be that the thick borders were from a single spammer, and they've > changed their design a little.
The test runs are on the same exact corpus. Comparing results for an older corpus and a newer corpus would make it impossible to find real problems. Daniel -- Daniel Quinlan anti-spam (SpamAssassin), Linux, http://www.pathname.com/~quinlan/ and open source consulting
