[WEB4LIB] Re: spam on staff computer - Korean

[Thomas Dowling]

At 01:45 PM 12/27/2002 -0800, Phalbe Henriksen wrote:
>Someone said they were filtering all of the Korean domains they could. I
>have received spam from Korea and because I downloaded the language support
>from Microsoft, I can read parts of the message, which are in English.
>
>The address it came from is <7love77 at hanmir.com>, if anyone wants to add
>that to their filters.


A serious spam problem from a single domain might be worth a trip to your 
mail server's sysadmin (armed with brownies, of course).  He or she can 
probably zap that domain's mail as it comes into the server, without it 
ever dropping those unneeded bytes into your account's spool file.  Even if 
spammers are too savvy to get caught this way, it can cut down on virus 
flare-ups coming from one infected computer out there somewhere.

I posted an earlier note about the spam analyzer in Mozilla 1.3a.  I've 
tried it for about a week now, and it seems to work as advertised.  Give it 
a set of messages with the appropriate ones labeled as junk mail and it 
learns on its own how to identify future junk mail.  Give it additional 
messages with junk and non-junk mail identified, and it fine tunes its own 
rules.  No need--well, *less* need--to create dozens or hundreds of ad hoc 
filters (can you even hope to write a filter that catches all variations of 
the "Nigerian Oil" fraud scheme?).

As an alpha version, Mozilla 1.3a isn't ready for primetime yet, but I'd 
bet cash money that this sort of analysis becomes the primary way of 
dealing with spam within the next year or two.


Thomas Dowling
Ohio Library and Information Network
tdowling at ohiolink.edu