Spam on the list [Was: Re: [R] Urgent ...]

Corrin Lakeland lakeland at lakeland.hn.org
Mon Nov 4 05:54:05 CET 2002


	    [MM: this message itself was also held as potential spam]

On Mon, 04 Nov 2002 17:19, you wrote:

> |> I'd hazard a
> |> guess that installing spamassassin would probably cause fewer
> |> problems.
>
> I, for one, am not impressed with spamassassin's algorithm for
> detecting spam.  It would have given the message I'm responding to
> points because it has a line all in capitals (the first one I've
> quoted).

>From that email:

X-Spam-Status: No, hits=-7.3 required=5.0
        tests=EMAIL_ATTRIBUTION,IN_REP_TO,PGP_SIGNATURE,
              QUOTED_EMAIL_TEXT,REFERENCES,SPAM_PHRASE_01_02,
              SUBJECT_IS_LIST,USER_AGENT,USER_AGENT_KMAIL
        version=2.43

I.e. the quoted email is very, very far from being spam.  

> Another one I've had experience with took exception to the
> fact that my Steven Wright quote contained the words "world's
> largest".

It seems you're not the only one, the line now reads:

/Join the world`s largest e-mail service with MSN Hotmail\./
	       ^ [changed by MM]
And incidentially, that line is considered evidence the mail isn't spam.

> Trying to make an algorithm that detects spam is kind of like trying
> to travel at the speed of light.  As soon as a system is set up, the
> spammers modify their tactics to "comply".

Bayesian based filtering? It is pretty hard for spam to keep up with 
something that changes every single message...

> The eye is a better judge of spam than a computer algorithm. 

And hence, Vipul's Razor?

>  I never
> read the spam in question because I automatically deleted it since I
> noticed it had the word 'urgent' in its subject.  

score URGENT 5.0?

I think you should give anti spam software a second try.

> Of course, if someone doesn't mind trawling through the rejects thrown
> out by a filter and checks that they really are spam, it would be
> greatly appreciated by the rest of the list 

This is a fair point.  One thing debian is thinking of doing is running 
a pyzor server.  Any user can report list spam to this server, and the 
other users won't see it.  Personally I don't like that idea here since 
it requires client-side software.

Corrin

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._



More information about the R-help mailing list