Re: should this all.SCORE file score my followups?

[Russ Allbery <eagle@cyclone.stanford.edu>]

Brian Edmonds <edmonds@cs.ubc.ca> writes:

> Not unless he wants billions of little ADAPT files all over.  Scoring on
> references isn't quite as accurate, but it's close and the overhead is
> quite a bit lower.  The main change he needs to make is 's' instead of
> 'e', as an exact match on his domain won't come up often...

The other problem he'll run into just scoring on host name in the
references line is that it will raise score on any articles from anyone
else at the same host.  If that's not the desired behavior, try scoring on
something like:

("references"
 ("\\.fsf\\(_-_\\)?@host\\.dom>" 200 nil r))

which worked for me until I managed to convince a bunch of other people
here to start using Gnus, at which point I made a copy of the message ID
generation function and changed it to put .thoron in the message ID
instead of .fsf.

Scoring based on references has saved me more time than any other single
feature of Gnus with the possible exception of mail splitting.

-- 
Russ Allbery (rra@cs.stanford.edu)      <URL:http://www.eyrie.org/~eagle/>