Al Iverson wrote:
On Mon, Apr 7, 2008 at 1:43 PM, Chris Lewis <clewis(_at_)nortel(_dot_)com>
wrote:
Very excellent job, Chris.
Did I hit the mark on collateral damage Rich?
Re, this bit:
Inevitably, some of these listings may impact non-abusive email.
This has, perhaps inevitably, resulted in some labelling such
practises by the emotionally loaded term "collateral damage". In
this context it has to be remembered that no filtering technique is
perfect, and that occasional mistake is inevitable no matter what is
used, DNSBLs or otherwise.
I have no idea off the top of my head for better wording, but I do
worry that the wording relating to mistakes implies that collateral
damage listings are never intentional; always in error. Besides the
potential for untruth there, it gives a listee a huge wedge to use to
complain about the list. "BCP says these kind of listings are
mistakes....and the list won't fix the mistakes!"
I'm fine with it if everyone else is, but I'm curious if others see
this as a potential issue as well.
Best,
Al Iverson
I've got a mixed opinion here, I'd break that down to say that
occasional mistake refers to the filtering method (And strictly to the
e-mail admin's configuration blocking e-mail that's not spam based on
DNSBL data. This is however a strict and very semantical reading of
that paragraph.
What about:
DNSBL policies vary on how they list and escalate filtering sources of
abusive content. This has in the past resulted in email which is not
spam being blocked by mail servers based on DNSBL input. This can occur
in two forms.
1. a DNSBL lists a "grey" source which emits both abusive, and
non-abusive e-mail which cannot be differentiated by DNS or IP.
2. a DNSBL as part of an "escalation" policy may list blocks of IP space.
DNSBL's SHOULD in the second case disclose and segregate in their
policies the specific policies and response codes which will contain
such blocks of IP space. This would include regional lists, RIR lists,
dial up user (DUL) lists, and listings of providers which have ongoing
or widespread abuse problems. This practice has become labeled by the
loaded term "collateral damage", but it's use is discouraged.
(NB I did not follow the argument on the use of "collateral damage" in
the long-winded thread on it's validity, so if the last sentence breaks
public consensus I am not looking to in any way re-open this argument.
Please modify it as appropriate and assume good faith).
Andrew
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/asrg