From: greg andruk (meowing@banet.net)
Date: Thu Jul 08 1999 - 10:02:09 CDT
In nnmh:usefor, Brad Templeton <brad@templetons.com> writes:
> Any commentary or X-trace-1 or whatever is not going to be of any use
> to outside spam filters etc.
Of course it can be used for filtering, /if/ the standard recognizes
that secondary injection will happen and provides a consistent way of
recording it. (Just to be clear here, I'm using "secondary" to mean a
proxy, or the on-the-train portaswerver in the running example.)
Even if the standard doesn't provide for a mechanism, reinjection
*will* still happen. Secondary injectors will simply continue doing
what they do now: invent custom headers of use to nobody else because
they're different in every implementation; or worse, simply strip off
the headers the primary injector doesn't like before posting and not
bother to note what was stripped.
The standard could say that secondary injection MAY be allowed, and if
it's allowed it MUST preserve certain headers before writing its own,
and the secondary injector MUST do the same things a primary injector
has to do. Then, no one has to support it but if they do, it can be
be done without the current uncertainty.