scoring help

scoring help

[Robert D. Crawford]

I am probably making this a lot harder than it need be, but I would like
for somebody to tell me that I have my head wrapped around this the
right way.

I would like to use adaptive scoring in nntp groups only.  I really only
want to set things up so that if I kill a message, I do not see the
followups.  There are a few more things that I have set, but you will
see those below.

First, I need to set this in my .gnus.el file

(setq gnus-default-adaptive-score-alist
       '((gnus-unread-mark)
         (gnus-ticked-mark (followup 5))
         (gnus-dormant-mark (followup 5))
         (gnus-del-mark (followup -5)
         (gnus-read-mark followup 5))
         (gnus-expirable-mark (followup -5))
         (gnus-killed-mark (followup -5))
         (gnus-kill-file-mark)
         (gnus-ancient-mark)
         (gnus-low-score-mark)
         (gnus-catchup-mark)))

So, any articles that I kill will get their score reduced by 5, and
anything I read, mark as dormant or tick will be increased by 5.

(setq gnus-score-exact-adapt-limit nil)

This means that only exact matching is done on the subject.

(setq gnus-summary-mark-below -5)

This will mark as read any articles where the score drops below 5.
True?  This means I will not see them?

(setq gnus-save-score t)

Necessary if I ever manually (in|de)crease scores.

(setq gnus-score-thread-simplify t)

this means that subject matching is done the same way as it is done for
threading. 

(add-hook 'message-sent-hook 'gnus-score-followup-thread)
(add-hook 'message-sent-hook 'gnus-score-followup-article)

(add-hook 'gnus-summary-mode-hook
   (function (lambda ()
     (gnus-summary-expunge-below -1000))))

this will expunge all messages from the summary that fall below -1000 in 
score.  I am still not sure if this is useful in nntp groups.

in the score file

      (mark -5)

wouldn't this mean that I will not see the articles?  Normally, when I
exit a group and go back to it, I do not see the articles that
have been read.  Does this change if using adaptive scoring?

      (expunge -1000)

In this case, does expunge mean the same thing as what happens when
you delete an article?  It is gone, never to return, it is an
ex-article.  I assume that this does not mark the article as read,
considering the mark-and-expunge variable below.  What is the
difference and when should it be used?  Also, is this not superfluous
with the gnus-summary-expunge-below variable?

      (mark-and-expunge -10)

See above.

      (read-only nil)
      (orphan -10)

So, if we expunge an article, and another article shows up in the
thread, the new article is an orphan, right?  This brings up
another question.  If we are talking about a newsgroup, is there any
need to expunge?  I was under the impression that setting things like
expiry in a nntp group was not necessary, since expiry was handled on
the server.  Am I correct?

      (adapt t)

This is pretty obvious.  I do have one question, though.  The variable
gnus-use-adaptive-scoring can be set to line or word, in addition to t
and nil.  If set globally to nil, and set to t in this group, does it
do scoring on line or word?  Since I have gnus-score-exact-adapt-limit
set to nil it really shouldn't matter, right?

I know this is long.  If you read this far, thanks.  Any help will be
appreciated.  

rdc

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Robert D. Crawford                                      rdc1x@comcast.net

Be security conscious -- National defense is at stake.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Re: scoring help

[Lars Magne Ingebrigtsen]

"Robert D. Crawford" <rdc1x@comcast.net> writes:

> This is pretty obvious.  I do have one question, though.  The variable
> gnus-use-adaptive-scoring can be set to line or word, in addition to t
> and nil.  If set globally to nil, and set to t in this group, does it
> do scoring on line or word?  Since I have gnus-score-exact-adapt-limit
> set to nil it really shouldn't matter, right?

Uhm...  probably.  :-)

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen