in message
Is there any way to block emails with foreign language text in
Outlook 2002?
Most of my spam is Russian, I think. I also have Norton Anti-Spam as
part of
the Internet Security package but I don't see how to block foreign
language
text.
You could define rules to look for the charset declaration. The
headers will contain something like:
Content-Type: text/plain;
charset="us-ascii"
In the above example, the e-mail was sent in plain text. The quoted
value is the name of the character set. You could either filter-in
which charsets you are willing to receive or filter-out those charsets
that you don't want to receive. Personally I would filter-out the
"bad" charsets. Some strings you could test on in the headers would
be:
charset="Big5"
charset="ChineseBig_Charset"
charset="EUC-KR"
charset="GB2312"
charset="GB2312_Charset"
charset="ISO-2022-JP"
charset="ISO-2022-KR"
charset="KOI8-R"
charset="KS_C_5601_1987"
charset="Windows-1250"
charset="Windows-1251"
charset="Windows-1254"
charset="Windows-1256"
charset="Windows-1257"
charset="Windows-1258"
charset="Windows-874"
You don't just define a rule to search the headers for 'Windows-1250'
(although it might work) but include the 'charset=' portion since it
is a parameter, usually on a separate line, of the Content-Type
declaration, so the rule would check headers for the presence of, for
example:
'charset="Windows-1250"'
You could have one rule that tests on all these charsets because the
operands gets OR'ed. I have the rule move these non-English e-mails
into the Junk folder which has auto-archiving enabled to permanently
delete items that are over 1 week old (for my personal/home e-mails;
at work, auto-archive to permanently delete the Junk folder is every 3
days since I am much more active in that instance of Outlook).
For rules that move e-mails into the Junk folder, I assign them to a
category. In the view of the Junk folder, I added the Category
column. That way, I can see why a message got moved into the Junk
folder. I also have multiple accounts in Outlook so I added the
E-mail Account column in the Junk folder (and most of my other
folders) so I can tell through which account an e-mail got received.
Personally I've found that using SpamPal with its IP backlists of
known spam sources to be a better gauge of what is spam or ham. Be
careful which blacklists you pick. I use SpamHaus SBL+XBL but not
PBL, and SpamCop. The Bayesian plug-in identifies spam whose source
hasn't been reported yet. I use the HTML-Modify but reconfigured it
to mostly provide added protection to OL2002. It wasn't until OL2003
when the option "Blocked external content" (i.e., linked images, like
web beacons) got added. HTML-Modify can nullify linked images. It
changes the "src=<url>" parameter to "xsrc=<url>" which is an illegal
parameter so the image doesn't get yanked but you could look at the
HTML code if you really needed to get that image. The UserLogf (user
logfile) plug-in saves a plain-text version of any spam-tagged e-mails
so you can go read those copies in case there was a false positive or
you want to report those e-mails later (because your rules in your
e-mail client [permanently] deleted them). SpamPal doesn't take any
action on spam. It only tags them. You then choose what to do with
the suspect e-mails based on what tags were added by SpamPal
(typically users just delete them). SpamPal and all plug-ins are free
and beat many commercial products.
I've tried other anti-spam products (free ones, that is).
Spamihilator is okay but it requires a ton of configuration and test
to reconfigure to eliminate the initially very high number of false
positives. It has taken me 3 man-days to start getting Spamihilator
to be useful without being overly aggressive. Their Spam Words filter
is worthless and results in a very high false positive (non-spam
tagged as spam) count. Some plug-ins are worthless, won't work, and
the vast majority of them have no documentation, and some don't even
have any configuration although they obviously need it. There are no
configuration settings for their Bayesian filtering (Learning Filter)
as there are with the Bayesian plug-in for SpamPal.
Using SpamPal or any other decent anti-spam program to filter out the
crap will work better than having to define a bunch of rules in an
e-mail client. I only have 2 rules defined in Outlook regarding spam
detection (a 3rd rule is to handle the spam-tagged e-mails by
SpamPal). My other 10 rules are not to detect spam but behavior in
some e-mails that I refuse to accept (blank subject, non-delivery
reports, no "@" in the From header, me in the From header, suspect
files but only if inline rather than attached, move to junk if I'm not
in the To/Cc headers, and the charset test, but not in that order). I
have SpamPal do the work of determining if an e-mail is spam or ham.
Trying to define your own rules means you will continuing be aiming at
a moving target and having to tweak your old rules or create new ones,
and I don't want to waste that much time on blocking spam. And NEVER
use the Blocked Senders list to eliminate spam since it won't work
against spammers that change their bogus e-mail address on every spew
of their crap.
