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

Re: poll: managing high volumes of list mail effectively

On Tue, 19 Oct 2004 05:36:04 +0200, martin f krafft <madduck@debian.org> said: 

> also sprach Manoj Srivastava <srivasta@ieee.org> [2004.10.17.0837
> +0200]:
>> Gnus can handle all that,

> Nice to know, but I ain't switchin' operating system. ;^>

>> Additionally, it has an adaptive scoring mechanism, that learns
>> about topics you are interested in, and topics and posters that one
>> rarely reads can progressively get even lower score.

> Interesting. mutt also features scoring, but I am not sure it's
> adaptive. How does Gnus do it? Sounds to me somewhat like a Bayesian
> module, or crm114 built in?

	Nothing so complicated.  In addition to static SCORE files
 (one per group/mailinglist, where my hand crafted rules (increase
 author substring permanent) go in), it has a per group/mailinglist
 ADAPT file as well, where automatically generated rules go.  The
 rules are generated based on actions I take on articles.

	here are the rules I use for actions (I may read an article,
 delete it, kill the thread, tick it to mark it persistent, etc). Each
 action marks the article in some way. For each mark, I specify scores
 to be added for various headers (usually from , the subject, or
 followup headers).

	Sounds complicated, but it isn't really -- I just deal with
 articles as I would, and slowly Gnus adapts to my likes.


;; for general rules
(setq gnus-default-adaptive-score-alist
        (gnus-ticked-mark      (from 256)   (subject 512)  (followup 1024))
        (gnus-dormant-mark     (from 512)   (subject 1024))
        (gnus-del-mark                      (subject -512))
        (gnus-read-mark        (from 32)    (subject 256))
        (gnus-killed-mark     (from -64)    (subject -512) (followup -1024))
        (gnus-kill-file-mark   (from -256)  (subject -128))
        (gnus-catchup-mark     (from -32)   (subject -256))))

;; for subject word scoring
(setq gnus-default-adaptive-word-score-alist
      '((,gnus-read-mark . 300)
        (,gnus-ticked-mark . 600)
        (,gnus-dormant-mark . 800)
        (,gnus-catchup-mark . -100)
        (,gnus-killed-mark . -200)
        (,gnus-del-mark . -150)))

Dijkstra probably hates me. Linus Torvalds, in kernel/sched.c
Manoj Srivastava     <srivasta@acm.org>    <http://www.golden-gryphon.com/>
1024D/BF24424C print 4966 F272 D093 B493 410B  924B 21BA DABB BF24 424C

Reply to: