[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

SLRN killfiles (was: pine & mutt)

Hash: SHA1

On Sun, 21 Feb 1999 21:13:38 -0700 (MST), John Galt wrote:

>I dunno about scoring--I rarely use the function, but I re-migrated back
>to TRN when I realized that SLRN had no killfiles, what I consider to be
>an essential tool in USENET.

    Uhm...  Uhm....  John, John, John....  *sigh*
    In the scale of how things work, scoring is *above* killfiles.
Killfiles are an absolute yes/no.  Do you get the article, yes or no?
Scoring are a set of rules that apply a positive or negative score to an
article.  Once all rules are applied to the article then action is taken on
that article, *including killing*.  A score of -9999 in SLRN's scoring *is*
functionally identical to a killfile.  

    For example:
Score:: -9999
Xref: .*:.* .*:.* .*:.* .*:.* .*:.*
Xref: advocacy
Subject: \$\$
Subject: \#\#
Subject: \<[Ss][Ee][Xx]\>
Subject: \!
Subject: 011
Subject: (?\d\d\d)?[- ]\d\d\d-\d\d\d\d
From: ;

    That is my "killfile", or a portion of it, really.  I didn't show and of
the "From:" filters.  But anything crossposted a few times, crossposted to
an advocacy newsgroup, that has $$, ##, a variation of sex, !, 011, a phone
number, or a ; in the related fields gets killed.

    But here is where killfiles get useful.  Let me show you three score
rules and an example of what can happen with them.  

Score:: 100
From: wardell@ibm\.net
From: davis@space\.mit\.edu

Score:: 100
Subject: [Jj][Oo][Ee]
Subject: [Jj][Ee][Dd]

Score: 1
~Subject: Re:

     Those three read that anything from wardell@ibm.net or
davis@space.mit.edu should be give a score of 100.  Anything with joe or jed
in the subject gets a score of 100, and anything that doesn't have Re: in
the subject is given a score of 1.  The first two rules are there because
they are people I want to read messages from, the last one gives a minute
boost to all new articles so when I go into a rather large newsgroup I can
skim the new headers and ignore the unscored articles, if I choose, because,
in theory, I've looked at the thread and deemed them not worth my time.

    So, if I go into comp.editors and John Davis, author of JED, is talking
about JED in a new thread, it gets a score of 201 and is filtered to the top
of my article list.  Any followups by other people are attached to it by
threading (for that session) and get a score of 100.  Any new articles, or
articles which contain joe/jed in the header get respective scores.  The end
result is an article list which is presented to me in a threaded format but
with articles I want to read near or at the top of the list.

    My favorite rule is this one:

Score:: 1000
References: morpheus@.*calweb.com
References: morpheus@.*rpglink.com

    Anything to me gets a score of 1000.  Shoots it right to the top of the
list.  I don't think I've ever missed a response to me.  Also, once I
participate in a thread the theory is that I am interested in it, so even
though the "1" score for "new thread" doesn't apply to later replies, I
still read these threads fairly quickly.

    Finally, this rule:
Score:: -500
From: vlouie@acs\.ryerson\.ca
From: wjr-kb@worldnet\.att\.net

    This is an individual I don't like reading messages from.  However, I
don't want to kill his posts because sometimes they may be directly to me
and I want to answer.  IE, he hasn't warrented the absolute "kill" of a
killfile.  So if he starts or participtes in a thread, that thread is moved
almost unilaterally to the bottom of my article list.  If he responds to any
post by me, the value is cut in half, placing it at the bottom of articles
at the top which are scored high for me.  IE, I get to go through all the
articles and then am presented with the "rubbish" that I can either choose
to read or discard, but has been filtered low for me so I know it is there.

    These are, admittedly, simplistic examples of what scoring can do.  You
can, obviously, score up and down on a variety of criteria that gives you
quite a diverse set of scores, both positive and negative, so that articles
are presented in a priority of your choosing.

    As for killfiles, as has been stated, anything -9999 is killed, no
questions asked.  Even messages which are to me (+1000) are killed because
the value -9999 is special.  So a +1000 and a -9999 is still -9999 and
killed.  The only way to prevent that is a +9999 *before* the -9999.  +9999,
as you may surmise, is "keep, no matter what."

    Another thing you can do, which is something I do not do, is set the
"kill" value lower than -9999.  You can set it to anything you want,
positive or negative.  Anything less than that value is killed.  So if you
have a bunch of negative rules that cumulatively get over that threshold,
boom, it is outta there.

    Hope this clears up the grave misunderstanding you, and others might,
have about SLRN.  :)

- -- 
         Steve C. Lamb         | I'm your priest, I'm your shrink, I'm your
         ICQ: 5107343          | main connection to the switchboard of souls.
- -------------------------------+---------------------------------------------
Version: PGPsdk version 1.0 (C) 1997 Pretty Good Privacy, Inc


Reply to: