Re: [scoring] How to filter trolls, kooks and fools who morph?

On 2006-03-15, Peter J Ross <pjr@xxxxxxxxxxxxxxx> wrote:

snip

What is your general filtering strategy? Do you plonk the From line
every time the target morphs, or do you look for patterns in other
headers?

Varies ;))

I'm not clever with regexs, but combining minor but consistent fragments
from different headers into single or multiple scorefile entries in slrn
is pretty flexible.

Some morphers seem to deliberately keep some elements of their From and
Message-ID headers consistent, perhaps so that they can easily track the
results of their own efforts - the trick then is to spot those
consistencies and devise ways of scoring down without getting too many
false positives. That can mean scoring up of other posters.

There don't seem to be any particular rules that can be set up in advance,
though.

If you've tried scoring on such "expensive" headers as Path
and NNTP-Posting-Host, does it slow down your Usenet reading too much?

I use leafnode to kill on:-

- crossposts to certain groups
- Followup-To to certain newgroups
- certain character sequences in the origination end of Path:

These settings tend to become permanent, and can make a great improvement.
It probably slows down leafnode's pulling of articles, but that seems to
be a price worth paying - and of course has no effect on slrn's
performance.

Sometimes I put an NNTP-Posting-Host in there for a while too.

Do you have any particularly clever regexes that catch morphers
unawares?

snip

No. One tip for slrn, though, is that you can put a ~ at the start of a
line in a score entry, to make it into a NOT test. One of these days I'll
find a use for that ;))

--
-- ^^^^^^^^^^
-- Whiskers
-- ~~~~~~~~~~
.

Relevant Pages

  • Re: c interview
    ... undefined behaviour because printf is a varidac function. ... stdio.h before using printf and other headers as appropriate before ... Would make it a point that the size of int is assumption to be .... ...
    (comp.lang.c)
  • Re: A critique of the BBC aquatic ape programme and the transcript.
    ... > Many, but not all, anatomical changes which are consistent with upright ... to be upright and move bipedally like no other factor. ... notice though that my criticism is not alone and it is not unique. ...
    (sci.anthropology.paleo)
  • Re: Question about the C-Reference manual
    ... >>At the bottom of page 254 there is the following code fragment: ... >>above gives the impression that even when an expression is undefined ... > 3) always evaluate parameters in some other 'consistent' order. ...
    (comp.lang.c.moderated)
  • Re: Idaho GOP leader has no concealed weapons permit
    ... propensity for various liberal reporters to try to put a "slant" on ... remote location as a political justification based on whatever ... Isn't 8 years consistent enough for you? ...
    (soc.retirement)
  • Re: c++ calling c functions
    ... CBFalconer writes: ... snip ... ... the headers. ... You can't invert this and generate normal C++ ...
    (comp.lang.c)