"Jeff Funk" <[EMAIL PROTECTED]> writes:

> So what does the average person set for the threshold of largest
> message scanned by SA.  Right now I'm at 32768 bytes.  How large is it
> safe to go???

By "safe", though, it's a bit unclear what you mean, but this should
help.  Here's my spam corpus broken down by size in bytes.

  count   bytes
--------------------------
      1   524288 - 2097152
      5   262144 -  524288
      3   131072 -  262144
     92    65536 -  131072
    152    32768 -   65536
    589    16384 -   32768
   1974     8192 -   16384
   5048     4096 -    8192
   7063     2048 -    4096
   3862     1024 -    2048
    273      512 -    1024
      2        1 -     512
--------------------------
  19064   all sizes

There is relatively more ham in the 512-1024 range and much more ham at
sizes larger than 16384 bytes.  Using large size as a rule never really
worked all that well, though.

256k seems like a good limit.  128k would be quite good too.

Daniel


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to