Ergo, facilitating heuristics is bad?
I'm not following the logic, but it must just be me. I don't see how
or why facilitating heuristics is bad -- for email, source IP
addresses, Subject line content, mismatched rfc821.from/rfc822.from,
body content ("Viagra", image/jpg, image/gif), are all useful as input
to a heuristic function.
I believe you're arguing that the sole purpose of duplicating header
information is to facilitate heuristics for use by a SpamAssassin-like
function, and that such duplication of header information has no other
purpose. Is that correct?
-d
On Nov 30, 2004, at 7:10 AM, Dave Crocker wrote:
On Mon, 29 Nov 2004 10:23:20 -0800, Dan Wing wrote:
With that in mind, I don't see arguments against SpamAssassin's use
of
heuristics apply to mailsig -- I don't believe anyone is proposing
or
suggesting that mailsig use heuristics. The fact that SpamAssassin
does so, in order to perform an additional function, is necessary
because of the additional function that SpamAssassin performs.
1. I think it's dandy for spamassassin to use heuristtics. In fact, I
think it's essential. However, I made a distinction between software
using heuristics, vs. standards.
2. I took Justin's note as providing encouragement to have the
standard duplicate headers, because it would facilitate some
heuristics.
So,
Having this be a justification for duplication of data is very much a
case of having mailsig participate in the use of heuristics, even
though the mailsig spec, itself, would not contain heuristics.
d/
--
Dave Crocker
Brandenburg InternetWorking
+1.408.246.8253
dcrocker a t ...
www.brandenburg.com