Stewart Stremler <[EMAIL PROTECTED]> writes:
> I don't know much about Gnus. How does the score affect what you
> see?
There are some significant score ranges -- configurable, of course --
that first give a message a classification as having either "high",
"default", or "low" score. By default, anything less than zero is low,
and anything greater than zero is high, with a configurable "fuzz"
range. Based upon this score, Gnus decides what treatment is
appropriate for the message. In one of my scoring configuration files
I have, among others, these rules:¹
(mark -175)
(expunge -400)
(mark-and-expunge -250)
The first says to mark any message with a score lower than -175 as
read. It's still present in the message list ("Summary Buffer"), but
it's marked as read (with an indication that the mark was automatic,
not the result of me having read it).
The third rule says to mark and remove any messages with a score lower
than -250. These messages will not be visible in the message list
unless explicitly requested by lowering the visible score threshold.
The second rule says to remove any messages with a score lower than
-400. We don't even bother to mark them as read, assuming we'll never
look at them.
Assuming that a given message is visible in the Summary Buffer, one
can tune the font used based on its score. The "Summary Buffer"
message list's format is configurable², of course, so one can display
the score value or use characters to indicate significant score
deviation. The default Summary format uses '+' and '-' to indicate
high- and low-scored messages.
The scoring system is rich. There's a whole Adaptive Scoring facility
that adjust message scores based on one's behavior.³ For instance,
reading a message may increase its score; marking it read without
reading it may decrease its score, as well as affecting the score
rules associated with the message's author or perhaps its containing
thread.
The best aggregate effect of scoring is cutting down on the number of
messages one faces when entering a given "group" (Gnus parlance for a
newsgroup or mail folder or something other collection of
messages). When I don't like the direction a thread is headed, I'll
hit, say
C-u 200 L R
to lower the score by 200 on all follow-up messages to the one under
consideration. You can see that -200 is near the mark-and-expunge
threshold of -250 mentioned above. If there's also a rule punishing a
given author that applies to a follow-up message, it may be enough to
drop that message entirely. Or as I keep reading an already suspicious
thread, I may add more score rules to prune follow-ups, but even then
I may allow tolerance for messages to survive from authors I'm
strongly interested in reading.
The default score delta for interactive rules is 1000. That means that
if I see a subject that I don't want to see again for a while, I can
just hit
L S
to lower the score by 1000 on all messages with that same subject
(substring match). One can optionally tune the score delta, the match
criteria (exact, substring, fuzzy, regex), the header (and some other
message facets) to match against, and the permanence of the score
rule. By default, score rules are "temporary": they decay in magnitude
over time as they fail to apply to new messages.
> I've never used the scoring feature of mutt. I perhaps ought to look
> in to that.
Neither have I. It's probably not as complicated or powerful as the
Gnus system. As I only use mutt occasionally in addition to Gnus, much
like I use vi occasionally in addition to (X)Emacs, I haven't put in
as much effort to tune it to my satisfaction.
Footnotes:
¹ http://www.gnus.org/manual/gnus_238.html
² http://www.gnus.org/manual/gnus_49.html#SEC49
³ http://www.gnus.org/manual/gnus_240.html#SEC240
--
Steven E. Harris
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
