🚀 go-pugleaf

In article <A9nGc.9456$yy1.4423@newsread2.news.atl.earthlink.net>,
mbox204 <noisp94@yahoo.com> wrote:
>
>I am continually irritated when I frequent some of my favorite newsgroups,
>this being one of them, only to find it being spammed with porn related
>messages.  It would be nice to prevent those types of postings in NGs's like
>this.

Yes, it would. Many such articles actually have cancels issued for
them by anti-spam people but, although it's not hard to distinguish
legitimate cancels from others (at least, some heuristics seem to work
pretty well, according to friends who run news systems), many ISPs
just ignore the lot so we don't benefit from that help. However, even
on the client side I'd have thought there are things that can be done:
news is quite mail-like, and there are even many newsreading programs
that are also MUAs so, given how well-developed anti-spam mail stuff
is these days, I'd have assumed that it's not too hard to find such
stuff for news too. (For instance, can Gnus' spam.el be made to work
for news? Maybe I'll give it a try someday.)

Still, http://www.usenet2.org/ might be of interest if you're not
already aware of it. I don't know if it's become a bustling community
of interested people, or if it's going the way of AX.25.

-- Mark

"mbox204" <noisp94@yahoo.com> wrote in
news:A9nGc.9456$yy1.4423@newsread2.news.atl.earthlink.net: 

> I am continually irritated when I frequent some of my favorite
> newsgroups, this being one of them, only to find it being spammed with
> porn related messages.  It would be nice to prevent those types of
> postings in NGs's like this.
> 
> 

some newsreaders, such as the program I use, Xnews, are highly effective 
against unwanted topics.  I have, on some newsgroups, eliminated 95% of the 
traffic, simply because of troublesome people/topics. 

John L. Wilkerson Jr. wrote:
> 
> some newsreaders, such as the program I use, Xnews, are highly effective 
> against unwanted topics.  I have, on some newsgroups, eliminated 95% of the 
> traffic, simply because of troublesome people/topics. 

What feature of Xnews allows this?

Neb Okla <n_okla@hotmail.com> wrote in
news:WHDGc.20694$2T2.4297@fe2.columbus.rr.com: 

> John L. Wilkerson Jr. wrote:
>> 
>> some newsreaders, such as the program I use, Xnews, are highly
>> effective against unwanted topics.  I have, on some newsgroups,
>> eliminated 95% of the traffic, simply because of troublesome
>> people/topics. 
> 
> What feature of Xnews allows this?
> 

The Scorefile.

John L. Wilkerson Jr. wrote:

> Neb Okla <n_okla@hotmail.com> wrote in
> news:WHDGc.20694$2T2.4297@fe2.columbus.rr.com: 
> 
> 
>>John L. Wilkerson Jr. wrote:
>>
>>>some newsreaders, such as the program I use, Xnews, are highly
>>>effective against unwanted topics.  I have, on some newsgroups,
>>>eliminated 95% of the traffic, simply because of troublesome
>>>people/topics. 
>>
>>What feature of Xnews allows this?
>>
> 
> 
> The Scorefile.

Care to elaborate?

In article <xJTGc.21339$2T2.18339@fe2.columbus.rr.com>,
Neb Okla  <n_okla@hotmail.com> wrote:
>
>John L. Wilkerson Jr. wrote:
>
>> Neb Okla <n_okla@hotmail.com> wrote in
>> news:WHDGc.20694$2T2.4297@fe2.columbus.rr.com: 
(snip)
>>>What feature of Xnews allows this?
>> 
>> The Scorefile.
>
>Care to elaborate?

I don't know about Xnews, but it's probably along similar lines as 
Gnus', described at http://tinyurl.com/3cat9 and subsequent pages.

-- Mark

Mark Carroll wrote:
> In article <xJTGc.21339$2T2.18339@fe2.columbus.rr.com>,
> Neb Okla  <n_okla@hotmail.com> wrote:
> 
>>John L. Wilkerson Jr. wrote:
>>
>>
>>>Neb Okla <n_okla@hotmail.com> wrote in
>>>news:WHDGc.20694$2T2.4297@fe2.columbus.rr.com: 
> 
> (snip)
> 
>>>>What feature of Xnews allows this?
>>>
>>>The Scorefile.
>>
>>Care to elaborate?
> 
> 
> I don't know about Xnews, but it's probably along similar lines as 
> Gnus', described at http://tinyurl.com/3cat9 and subsequent pages.
> 
> -- Mark

Ok, so apparently it ranks based on keywords.  Seems similar to Bayesian 
filtering used by Outlook, Thunderbird, and others - except possibly 
requiring more maintenance.

So far I haven't found any anti-spam technique that impresses me, but I 
like the idea of bandwidth throttling for machines that are sending out 
"too much" email (until it happens to me of course).

In article <BiYGc.1184$wX6.43@fe1.columbus.rr.com>,
Neb Okla  <n_okla@hotmail.com> wrote:
(snip)
>So far I haven't found any anti-spam technique that impresses me, but I 
>like the idea of bandwidth throttling for machines that are sending out 
>"too much" email (until it happens to me of course).

I've yet to try it out properly, but one thing that does intrigue me
is collaborative spam detection. For example, Vipul's Razor claims to
establish a distributed and constantly updating catalog of spam based
on consensus judgments by users. So, enough people mark it as spam,
enough more than not-spam, then your system recognizes it as spam
automatically. Of course, this requires some network bandwidth for
talking to the distributed catalog, and presumably some cleverness
to protect against malicious users.

-- Mark

Mark Carroll wrote:
> 
> I've yet to try it out properly, but one thing that does intrigue me
> is collaborative spam detection. For example, Vipul's Razor claims to
> establish a distributed and constantly updating catalog of spam based
> on consensus judgments by users. So, enough people mark it as spam,
> enough more than not-spam, then your system recognizes it as spam
> automatically.

My only concern here would be that if I wanted to recieve email from 
Wal-Mart, Starbucks, or Microsoft (for example) some people might 
declare those as Spam just because they don't like the company.

While I think this should be possible on a personal level, I wouldn't 
want to rely on others accusations in all cases (though 99% of the time 
they'd probably be right).

Some strategies that interest me are "Penny Black" along with some 
others that I haven't seen discussed for implementation - like detecting 
   strings of random characters - or long lists of random phrases thrown 
in to fool Bayesian filters.

Ultimately it would seem that a combination of these strategies might 
offer the best protection, but ultimately the ability to reject email 
from forged addresses would probably weed out the most junk.

Of course, I'd also like to see Spam cut-off at the source as well.

Neb Okla <n_okla@hotmail.com> wrote in news:xJTGc.21339$2T2.18339
@fe2.columbus.rr.com:

> John L. Wilkerson Jr. wrote:
> 
>> 
>> The Scorefile.
> 
> Care to elaborate?
> 

A filtering system that allows filtering on many of the message headers.  
The file can be changed thru a text editor or to a lesser degree through 
the scoring menu.


Details are at the Xnews site:

http://xnews.newsguy.com/

Neb Okla <n_okla@hotmail.com> writes:

> Ok, so apparently it ranks based on keywords.  Seems similar to
> Bayesian filtering used by Outlook, Thunderbird, and others - except
> possibly requiring more maintenance.

Scoring is very effective, not just for dealing with spam, but for
being able to sort things based on likely interest.  I use "adaptive
scoring" in Gnus and am able to keep up with about 350 newsgroups,
another fifty to seventy-five mailing lists, and probably another 100
RSS feeds because of adaptive scoring.  People whose articles I read
and words in subjects of articles I frequently read get scored up, and
the opposite happens on articles on which I hit "d", as if I'm
deleting them even though I'm probably not actually doing that.

For the record, I issue third-party cancels on cmh.* newsgroups where
malware is attached, but I do this manually instead of with a program
that would issue them immediately.

I've been thinking about using NoCeM to adopt the hierarchy and then
issue NoCeM messages instead of actual cancels.

-- 
Matt Curtin, CISSP, IAM, INTP.  Keywords: Lisp, Unix, Internet, INFOSEC.
Founder, Interhack Corporation +1 614 545 HACK http://web.interhack.com/
Author of /Developing Trust: Online Privacy and Security/ (Apress, 2001)

In a recent article, I mentioned NoCeM and third-party cancels.  I got
some follow-up questions privately, and thought I'd post some
additional information about these.

First of all, I do not recommend issuing third-party cancels.  It's
one of those "if you don't know how to do it, you shouldn't be asking"
kinds of things.  If a Usenet Cabal existed, the Cabal would issue
cancels, but no one else should. N.B., There Is No Cabal.

Cancels in Usenet are a method by which an author may decide to delete
a message that he has posted.  Like any other message, the cancel will
travel throughout Usenet, except instead of adding a message to the
news servers' spools, it'll remove an identified message from the news
servers' spools.  Third-party cancels (i.e., one user canceling
another user's message) are possible because cancel messages have no
method for authenticating the cancel issuer.  Many sites ignore all
cancel messages (because they're easy to be abused), while others
accept them (because they're able to keep spam from knocking over
stressed Usenet servers).

NoCeM (pronounced, "no see 'em") is a method for identifying and
preventing spam from being seen.  Rather than actually removing
messages from spools (which is what happens with cancels), the
messages are still there, but people who wish to obey the NoCeM
messages of particular issuer may do so.  For example, if I start to
issue NoCeM messages, individual readers may decide whether they would
like automatically to accept my categorizations.  Users with NoCeM
capable clients configured to accept my NoCeM messages will then
simply not see the messages for which I issue a NoCeM.  Integrity and
authenticity of NoCeM messages is handled by PGP digital signatures.

More information is available from "Fight Spam on the Internet!"[1]
and CancelMoose[2].


Footnotes: 
[1]  http://spam.abuse.net/spam/

[2]  http://www.cm.org/

-- 
Matt Curtin, CISSP, IAM, INTP.  Keywords: Lisp, Unix, Internet, INFOSEC.
Founder, Interhack Corporation +1 614 545 HACK http://web.interhack.com/
Author of /Developing Trust: Online Privacy and Security/ (Apress, 2001)