administrivia: filtering

[Richard Welty <rwelty@xxxxxxxxxxxxxxxxxxxxxx]

since a lot of people have gotten resigned since i sent this out,

there is now extensive filtering on the bmw-digest. this stuff serves
several purposes: spam control, html/mime control, and "cooling off"
intemperate writers.

we are blocking html, ms-tnef, base 64 encoding, winmail.dat files, some
types of mime-encoded signature, and those silly virtual business card
files. if you think you might be sending anything of the sort, please turn
it off when sending to the digest. if you don't, it bounces into my inbox
where i will most probably send it back to you to be edited and resent,
sans the extraneous material.

additionally, i have developed a fairly extensive library of strings that
match common spams, and it improves each day -- unfortunately -- because of
the enormous volume of spam to which i personally am subjected because i've
been on the net since 1982, and every address i've used since 1985 still
forwards to me (and believe me, the spammers have found and are using
addresses that i've not been able to log onto since the late 80s. there is
one spam list for sale out there that has 3 or 4 different addresses that
forward to me on it. quick, now, how many of you have been around long
enough to remember when i was welty@domain.elided?)

anyway, the filtering is set up so that when certain regular expressions
match, the email gets put into my inbox for review. this means that if you
accidentally match a string, your posting might still get out. here are
some examples of things that we're matching in headers (the / characters
delimit the string; the i option after the last / means that the search is
case insensitive; the ^ matches beginning of line; \b matches arbitrary
inter-word breaks; .* matches an arbitrary number of characters)

/^subject:.*investment\bopportunity/i
/^subject:.*requested.*info/i
/^subject:.*info.*requested/i
/^subject:.*ADV:/i
/^subject:.*AD:/i
/^subject:.*cash/i
/^subject:.*golf/i
/^subject:.*make money fast/i
/^subject:.*income/i
/^subject:.*xxx/i
/^subject:.*adult/i
/^subject:.*sex/i
/^subject:.*Remember Me/i

if you match one of these, or any of a number of others, your mail will be
delayed.

here are some samples of expressions being matched against message bodies
(\d+ matches one or more digits; s? matchs 0 or 1 letter s; -? matches 0 or
1 hyphen):

/\bare\byou\bin\bdebt\b/i
/\bMLM\b/i
/\bbusiness\bopportunity\b/i
/\bcable\bdescrambler\b/i
/\bpheremone\b/i
/\d+\bcents?\bminute/i
/\bglobal\bremove\b/i
/\bthis\bmessage\bcomplies\b/i
/murkowski/i
/\bremoval\binstructions\b/i
/\bwish\bto\bbe\bremoved\b/i
/\bjunk\be-?mail\b/i
/\binternet.*merchants\b/i
/\badvertise.*internet\b/i
/\bbulk\be-?mail/i
/\bdirect\be-?mail\b/i
/free\bfree\bfree/i
/\bXXX\b/
/\bnude\b/i
/"remove" in the subject line/i
/"remove" in the subject field/i

so that's the story. the actual lists are more extensive, but now you all
get the picture about the filtering capabilities that come with the new
server.

cheers,
  richard

------------------------------