Automatic generation of split methods

Automatic generation of split methods

[Benny Amorsen]

It would be nice if gnus could automagically generate
nnmail-split-methods a la the automagical generation of score file
entries. Maybe make another file containing split methods.

M-x gnus-summary-generate-split-method-that-matches-the-author-of-this-article


Benny

-- 
Finger for PGP Public Key
WWW: <A HREF="http://www.daimi.aau.dk/~amorsen/">Benny's homepage</A>

Re: Automatic generation of split methods

[Lars Magne Ingebrigtsen]

Benny Amorsen <amorsen@daimi.aau.dk> writes:

> It would be nice if gnus could automagically generate
> nnmail-split-methods a la the automagical generation of score file
> entries. Maybe make another file containing split methods.
> 
> M-x gnus-summary-generate-split-method-that-matches-the-author-of-this-article

Hm.  I think creating split methods in this manner might be nice, but
Yet Another Startup File?  There are quite a lot of them already...

-- 
  "Yes.  The journey through the human heart 
     would have to wait until some other time."

Re: Automatic generation of split methods

[Mats G. Lofdahl]

Benny Amorsen <amorsen@daimi.aau.dk> writes:

> It would be nice if gnus could automagically generate
> nnmail-split-methods a la the automagical generation of score file
> entries. Maybe make another file containing split methods.

Why not implement splitting for the other backends as well?  Then put
split methods in the score files?  And merge splitting and scoring
interfaces?  

Splitting of nntp groups would make it possible to e.g. have subgroups
to high-volume groups only containing particular threads, threads with
contributions from particular authors (like yourself), announcements,
high scores, etc.  Wouldn't this make it faster to enter (sorting much
fewer threads) high volume groups for just looking at interesting
stuff, while retaining the possibility to look at the other threads
some other time (which you don't have if you score low enough to
kill).  

Or would this be incredibly slow?

-- 
    Mats Löfdahl                   Internet: Mats@Astro.su.se
    Stockholm Observatory                    MLofdahl@Solar.Stanford.edu
    S-133 36 Saltsjöbaden          Voice:   +46 - 8 16 44 75
    Sweden                         Fax:     +46 - 8 16 42 28

Re: Automatic generation of split methods

[Andy Eskilsson]

/ Mats.Lofdahl@astro.su.se (Mats G. Lofdahl) wrote:
| 
| Why not implement splitting for the other backends as well?  Then put
| split methods in the score files?  And merge splitting and scoring
| interfaces?  
| 

Yes! 

This would be wonderful, I can do nice stuff like slam together all
the win95 newsgroups (yes I know, I am a pervert!) into one big
virtual group, from wich I send all articles from Bill Gates into
comp.windows95.god, all threads about something like programming into
comp.windows95.hacking.. 

Guess this would beat the sh*t out of the old sun 1+ with 16 meg I am
sitting at, but howabout some kund of batch-mode that makes gnus do
all the heavy work in the background (collecting, scoring sorting for
all the groups) while I am working with something else in another
window? sounds like some kind of soupbrewing, but
brewing+scoring+sorting? 

	/Andy

Re: Automatic generation of split methods

[Per Abrahamsen]

You have just described the nnkiboze backend.

Re: Automatic generation of split methods

[Andy Eskilsson]

/ Per Abrahamsen <abraham@dina.kvl.dk> wrote:
| 
| You have just described the nnkiboze backend.
| 

OOps.. yes, but only almost I think.. does nnkiboze do scoring
threading and sorting too?

	/Andy

Re: Automatic generation of split methods

[Andy Eskilsson]

/ Mats.Lofdahl@astro.su.se (Mats G. Lofdahl) wrote:
| 
| Why not implement splitting for the other backends as well?  Then put
| split methods in the score files?  And merge splitting and scoring
| interfaces?  

It struck me, once in the old ancient ding gnus, I messed around with
some way to split a mailgroup with a regexp, and then switch between
the different 'partitions' in the *Summary buffer*.. Humm can't find
it now with a quick glance through the docs..

	/Andy