To be pedantic, there are the following alternatives:
1) Force all FAQs to be posted from the rtfm server. Ugh.
2) Ship a copy of the private key (ugh!) for the rtfm server
to every faq poster, and force every faq poster to run
the PGPMoose client (ugh). You'd need instantaneous
revocations/reissue of keys. No such thing yet.
3) force every faq poster to use PGP, deliver their public key
to every server, and use PGPMoose client. Again, no PKI is
commonly available. And PGPMoose code is nasty business
(I know, I helped Glen write it).
Something cheaper/more portable than PGPMoose would be nice.
} Marty> 2) exempt moderated groups from spam BI calculation, other than
} Marty> checking the signature.
} > This is a non-issue for FAQs. How often must we say this: *no
} > currently approved FAQ is breaking the spam thresholds*.
} Once was more than enough. I thought we had moved on to Chris' point
} that cancelling, in general, should not rely on content or intent of
} posting, but rather BI. My suggestion was pointed at that, with an
} intent that FAQs that might someday break the current BI would be
} covered as a result.
I'm not sure that this is desirable actually (entirely aside from
"content blindness" doctrine).
} > *No-one*, certainly not myself, Chris, or the usenet-format group, can
} > say to news admins "you must exempt FAQs from your server's filters".
} > We can work with the authors of those filters to try and ensure that
} > sensible defaults are used, and we can advise faq-maintainers as to
} > crossposting limits and maximum posting frequency.
} Nor did I anywhere say you could. I think I've wandered into
} someone's hot topic. My apologies.
Not really. We're just seeing a few people not quite understanding things,
and we (blindly) mistook you for one of them... Sorry Marty.
As someone else suggested, having the filters work on BI would be
nice. However, in order to do this in any semi-automated fashion,
you have to invest in a full blown "Cosmo Roadkill" engine. Aside
from the fact that Cosmo ain't offering it for use at the moment,
you'd be literally condemning every ISP to have to buy a rather largish
machine just to run it. Hippo, on the other hand, is nicely
"light weight".
We need to invest in education here somehow. Perhaps a "watch out
for filters!" faq ;-)
Someone else mentioned exempting moderated groups from BI calculations.
There are some possibilities here.
The policy with moderated groups is that articles in moderated groups
are subject to spam cancellation by default. [This has been discussed
several times on the moderator's list, and met with overwhelming support.]
Furthere, any "well moderated" (means: human moderated. Autoreflectors
don't count - even with spam filtering) group can elect to opt out of spam
cancellation by having a majority of the moderator[s] of record requesting
it.
To date, precisely three groups have done so. [I should point out that
"have done so" is with me. Undoubtably, some other spam cancellers will
not be aware of this, and are cancelling anyways. I find out about these
things since I robonotify spammers, and hence the moderator "sees" me
doing it.]
Problem is, however, that since an unmoderated group has no "moderator
of record", a posting cross-posted between a moderated and unmoderated group
is subject to spam cancellation regardless of what the moderator wants.
So, while Pam et. al. could formally request us to exempt *.answers,
it would only work for those FAQs _only_ in moderated groups.
Perhaps we could eliminate the mention of a moderated group from BI
calcs. So, a FAQ would have to be massively crossposted into unmoderated
groups to hit BI 20, but, it would still be cancelled everywhere if it did.
[It's not possible to cancel a crossposting out of just some of the
groups. It's all or nothing.]
-- Rumours of my demise are greatly exaggerated. Cold but happy, please pass the thermos!For more information on spam, see http://spam.abuse.net/spam Fight spam, support Rep. Chris Smith's TCPA extension: http://www.cauce.org