[leafnode-list] Re: Leafnode spam learning?

[Matthias Andree <matthias.andree@xxxxxx>]

"Carr, Chris" <Chris.Carr@xxxxxxxxxxxxx> writes:

> Hi All,
>
> I run Leafnode so that I can read news from anywhere I like - home,
> work, 
> e-cafe etc. This means I use several different clients: OE at work
> because I 
> can't install anything else, Thunderbird at home in Windows, Evolution
> in 
> Linux etc. etc.

Chris,

apparently there's something wrong with your line wrapping (looks
Microsoftish and comb-style, one long (the "prong") and one short (the
back of the comb) - try setting a very long line length in OE and wrap
manually when typing your text) - I've left it intact for the paragraph
quoted above for you to see and am fixing the other paragraph.

> Traditionally, spam-filtering has been seen as a client-side activity,
> but it would clearly result in much redundant work for me. I'm looking
> for a way of hooking up an intelligent spam filter (like SpamAssassin)
> with Leafnode, so that it regularly updates /etc/news/leafnode/filters
> for me.

I'm not sure if a modification of the .../filters file would be the
right approach, and I think that the leafnode regexp filtering is too
lightweight for someone to transform SpamAssassin rules to them.

I agree that server-side filtering would be useful here, because
client-side filtering doesn't work in many of the popular newsreaders or
combined mail/newsreaders for NEWS.

I HAVE however plans to provide some hooks for scripting in leafnode-2
later on. Ideally, these would then be amenable to hook up the usual
(spam) filter candidates such as bogofilter, crm114, qsf, spamassassin,
spamprobe or others (not sure if dspam is amenable here, I know it's
popular but don't know it), and perhaps also hook more generic article
schedulers, and perhaps the interface would allow reusing filter
software that was developed for news software (INN comes to mind).  Note
this is an ideal and as of yet unimplemented scenario, working in the
beautiful city of Utopia, I'm not yet sure which parts of it will later
actually be implementable for leafnode-2.

The problem with that is that I'm basically the only developer for
fetchmail and leafnode nowadays, and fetchmail is already behind.
So if someone has interest and spare time to help, let me know.

HTH

-- 
Matthias Andree
-- 
_______________________________________________
leafnode-list mailing list
leafnode-list@xxxxxxxxxxxxxxxxxxxxxxxxxxxx
https://www.dt.e-technik.uni-dortmund.de/mailman/listinfo/leafnode-list
http://leafnode.sourceforge.net/