Re: Amazon / RFCI false positives

[Justin Mason <jm...@jmason.org>]

jdow writes:
> From: "Chris Lear" <ch...@laculine.com>
> >* Tony Finch wrote (05/11/06 17:43):
> >> On Sat, 4 Nov 2006, Michael Scheidell wrote:
> >> 
> >>> So? Build something better. Its open source. Don't use the RFCI scores,
> >>> drop them, stop bithing about somehting YOU can change.
> >> 
> >> Well, I've added a -2 for email from Amazon, but I thought other people
> >> might like a warning.
> > 
> > Thanks. Warning appreciated.
> > 
> > I think that the people who made derogatory claims about "Tony's logic",
> > or claimed that "you don't understand" had failed to appreciate what
> > "These messages are wanted by their recipients so should not be
> > scored as spam by SpamAssassin" means. Anyone who disagrees with that
> > piece of logic would appear to be using Spamassassin for a purpose that
> > its designers didn't think of.
> 
> Tony's phrasing implied that he thought the scoring was so wrong
> that it should be modified by the people who wrote the rule and ran
> it against mass checks. That logic is dead wrong.
> 
> The correct phrasing might have indicated there is a problem for some
> sites with Amazon failing RFCi requiring a special rule to negate
> Amazon.com's negative scores on RFCi.
> 
> Demanding that the RFCi rules vanish into the night just is not going
> to fly. And it indicates flawed thought processes.

Well, some of them recently *have* -- they were just
becoming very poor indicators of spam/nonspam status, and that's
what SpamAssassin rules are used for.

--j.