[gentoo-dev] Spam (was Re: Hey SEXY)

[gentoo-dev] Spam (was Re: Hey SEXY)

[Henrik Treadup]

Alan wrote:
#
#Or filter the list through spamassassin before it goes out to mailman (or
#something like that).
#
#alan 


But spamassassin is sloooooow. It also uses a human generated rulebase.
http://www.spamassassin.org/tests.html
Lots of work for humans which is bad :( 


What we wan't is a rulebase specifically tailored to the gentoo lists. (If a 
mail contains the word 'ebuild' it is probably not spam and if it
contains the word 'madam' it probably is) 


Bayesian filtering takes care of generating the rulebase automatically. This 
is a Good Thing (TM). 


Read 'A plan for spam'  http://www.paulgraham.com/spam.html and you will see 
the light ;) 


Where would be the correct place to have a discussion about this BTW? 

/Henrik Treadup
hetr9922@student.su.se 

Re: [gentoo-dev] Spam (was Re: Hey SEXY)

[Matthew Walker]

Henrik Treadup said:
> Alan wrote:
> #
> #Or filter the list through spamassassin before it goes out to mailman (or
> #something like that).
> #
> #alan
>
>
> But spamassassin is sloooooow. It also uses a human generated rulebase.
> http://www.spamassassin.org/tests.html
> Lots of work for humans which is bad :(
>
>
> What we wan't is a rulebase specifically tailored to the gentoo lists. (If
> a  mail contains the word 'ebuild' it is probably not spam and if it
> contains the word 'madam' it probably is)
>
>
> Bayesian filtering takes care of generating the rulebase automatically.
> This  is a Good Thing (TM).

Bayesian Filtering /rocks/! I've written a Bayesian spam filter for
SquirrelMail based on the article you mention further down, and it's great.
I don't see the spam on mailing lists anymore, because the filter catches it
and filters it to another folder. Too bad I still get to see everyone
talking about how much spam sucks. ;)

>
>
> Read 'A plan for spam'  http://www.paulgraham.com/spam.html and you will
> see  the light ;)
>
>
> Where would be the correct place to have a discussion about this BTW?
>
> /Henrik Treadup
> hetr9922@student.su.se
> _______________________________________________
> gentoo-dev mailing list
> gentoo-dev@gentoo.org
> http://lists.gentoo.org/mailman/listinfo/gentoo-dev

[gentoo-dev] Spam (was Re: Hey SEXY)

[Henrik Treadup]

I don't like the idea of closing this list. If you close it you will still 
have problems with spam in the open lists; user, newbie etc. A better way of 
doing things would be to filter out the spam completley on all the lists. 

There was an interesting thread about Bayesian spam filtering on slashdot
http://slashdot.org/article.pl?sid=02/09/17/1746248&mode=thread&tid=111
a while ago discussing Paul Grahams essay A Plan for Spam
http://www.paulgraham.com/spam.html 

The  statistical approach described there would have filtered out all spam I 
have seen on this list so far. (How many non spam email contain the words 
'dear' and 'madam' ?) 

You might wan't to check out ESR's implementation 
http://bogofilter.sourceforge.net 

Keep up the good work
/Henrik Treadup
hetr9922@student.su.se 

PS. The logical name for the distribution is Gnutoo Linux  ;)

Re: [gentoo-dev] Spam (was Re: Hey SEXY)

[Alan]

On Fri, Sep 27, 2002 at 02:12:31AM +0200, Henrik Treadup wrote:
> I don't like the idea of closing this list. If you close it you will still 
> have problems with spam in the open lists; user, newbie etc. A better way 
> of doing things would be to filter out the spam completley on all the 
> lists. 
> There was an interesting thread about Bayesian spam filtering on slashdot
> http://slashdot.org/article.pl?sid=02/09/17/1746248&mode=thread&tid=111
> a while ago discussing Paul Grahams essay A Plan for Spam
> http://www.paulgraham.com/spam.html 
> 
> The  statistical approach described there would have filtered out all spam 
> I have seen on this list so far. (How many non spam email contain the words 
> 'dear' and 'madam' ?) 
> 
> You might wan't to check out ESR's implementation 
> http://bogofilter.sourceforge.net 

Or filter the list through spamassassin before it goes out to mailman (or
something like that).

alan

-- 
Alan "Arcterex" <alan@ufies.org>   -=][=-   http://arcterex.net
"I used to herd dairy cows. Now I herd lusers. Apart from the isolation, I
think I preferred the cows. They were better conversation, easier to milk, and
if they annoyed me enough, I could shoot them and eat them." -Rodger Donaldson

Re: [gentoo-dev] Spam (was Re: Hey SEXY)

[Mark Bainter]

Alan [alan@ufies.org] wrote:
> On Fri, Sep 27, 2002 at 02:12:31AM +0200, Henrik Treadup wrote:
> > I don't like the idea of closing this list. If you close it you will still 
> > have problems with spam in the open lists; user, newbie etc. A better way 
> > of doing things would be to filter out the spam completley on all the 
> > lists. 
> > There was an interesting thread about Bayesian spam filtering on slashdot
> > http://slashdot.org/article.pl?sid=02/09/17/1746248&mode=thread&tid=111
> > a while ago discussing Paul Grahams essay A Plan for Spam
> > http://www.paulgraham.com/spam.html 
> > 
> > The  statistical approach described there would have filtered out all spam 
> > I have seen on this list so far. (How many non spam email contain the words 
> > 'dear' and 'madam' ?) 
> > 
> > You might wan't to check out ESR's implementation 
> > http://bogofilter.sourceforge.net 
> 
> Or filter the list through spamassassin before it goes out to mailman (or
> something like that).

                    Close the list      Filter the lists
False Positives          No                  Yes
100% effective          Yes                  No

Hrm...choice seems clear to me.  Filtering spam on a
mailing list is a bad idea.  It's one thing for an 
individual user to decide an acceptable level of risk
for false positives, it's quite another to choose that 
for an entire list.  Just my two cents.

Re: [gentoo-dev] Spam (was Re: Hey SEXY)

[jsykari]

Mark Bainter <mark-gt@cymry.org> writes:
> Alan [alan@ufies.org] wrote:
>> Or filter the list through spamassassin before it goes out to mailman (or
>> something like that).
>
>                     Close the list      Filter the lists
> False Positives          No                  Yes
> 100% effective          Yes                  No
>
> Hrm...choice seems clear to me.  Filtering spam on a
> mailing list is a bad idea.  It's one thing for an 
> individual user to decide an acceptable level of risk
> for false positives, it's quite another to choose that 
> for an entire list.  Just my two cents.

How about making a new list, gentoo-dev-spam, and sending the filtered
mail to that address? Those who want to avoid false positives can
freely subscribe to that list.

This might seem ridiculous but it's not as ridiculous as doing nothing
and letting the mailing list get cluttered by spam. How much spam is
enough to make the list closed? 10 messages per day? 100 per day?

Better even, put the address "gentoo-dev-spam@gentoo.org" in the front
page of www.gentoo.org.  That would further raise the signal/noise
ratio of gentoo-dev, which is now, IMHO, rather much plagued by
messages which would rather belong to gentoo-newbie or
gentoo-user. (Or forums.gentoo.org)

To be a bit more realistic, a bit more realistic solution is to close
the lists.

Distributing spam on the mailing lists is increasingly hurting the
gentoo community by hindering communication. To not close the mailing
lists would be a sign of weakness and inability to make
decisions. Personal mail filters should not - and could not - be the
solution, because not everybody has the possibility, resources or
willingness to install SpamAssassin or similar.

Antti, fed up with spam

Re: [gentoo-dev] Spam (was Re: Hey SEXY)

[Moritz Schulte]

Mark Bainter <mark-gt@cymry.org> writes:

> Hrm...choice seems clear to me.  Filtering spam on a mailing list is
> a bad idea.  It's one thing for an individual user to decide an
> acceptable level of risk for false positives, it's quite another to
> choose that for an entire list.

Spam does not have to be filtered out in the sense of dropping the
mails.  Mails, which look like spam to a filter, could be extended
with an email header tag.  Then all mails would still be delivered,
but users are free to filter out those mails, which are tagged as
spam.

		moritz
-- 
moritz@duesseldorf.ccc.de - http://duesseldorf.ccc.de/~moritz/
GPG fingerprint = 3A14 3923 15BE FD57 FC06  B501 0841 2D7B 6F98 4199

Re: [gentoo-dev] Spam (was Re: Hey SEXY)

[Fredrik Jagenheim]

On Thu, Sep 26, 2002 at 05:17:53PM -0700, Alan wrote:
> Or filter the list through spamassassin before it goes out to mailman (or
> something like that).
> 

I'm running spamassassin localy, and while it catches most spam (way
more than my custom procmail filter did) some spam comes through.
This one (Hey SEXY) did, and a couple of others have too.

Close the lists, I see no reasons why not to. As I stated in another
letter.

Brgds,
//Humming