Bogofilter accuracy plummets starting around March 10, 2010

[Jonathan Kamens]

Thomas Anderson wrote:
> Also, in the end, spamitarium doesn't throw away any of the forged 
> received line info... So information was ADDED, not removed, 
> by spamitarium.
>   
I did a little playing around with spamitarium, and it appears that it 
/does /throw away useful information from /valid/ Received lines.  For 
example, I fed a recent valid message through spamitarium with this 
valid Received line in it:

Received: from omta01-mdp.westchester.pa.bo.comcast.net ([76.96.53.11])
      by qmta02-mdp.westchester.pa.bo.comcast.net with comcast
      id 1u2R1e0010EWbPw01u407n; Mon, 05 Apr 2010 18:04:00 +0000

and it spit back this:

Received: from omta01-mdp.west 76.96.53.11 as36733
        by qmta02-mdp.westchester.pa; Mon, 05 Apr 2010 18:04:00 +0000

So the host names in the received line are no longer FQDNs (in violation 
of RFCs, I believe) and the MTA queue ID has been discarded.

Both of these are of concern.  The latter is a problem if there is an 
issue with mail delivery and I need to contact a transit site about it 
(and yes, I've done this on numerous occasions in the past).  If I can 
give them the queue ID assigned to the message by their MTA, then they 
can find it easily in their logs.  If I can't, I might as well forget 
any possibility of them being willing to help me.

  jik