Interview with a Spammenter
Posted January 31st, 2005 @ 04:10pm by Erik J. Barzeski
I found this interview interesting. It reminded me to check my logs. Apparently I'm doing pretty well for myself with MT-Blacklist. In the past 30 days, I've had some 25,000 spamments blocked. I've had to manually delete only 7 spamments.
My blacklist makes liberal use of URL Patterns like "\bdebt\b" and "\bhold-em\b" and it's been working well. They catch nearly all of the comment spam. Some others - "phentermine" and "vioxx" and "cialis" (which, admittedly, will trip up people who spell "specialize" with an "s" instead of a "z") - act as good catch-alls.
Update: I've updated my blacklist message with to list the affected word, so if you've commented here before and been blocked, the message should be more helpful now. ("Your __SUBMTYPE__ was denied for questionable content. The item matched was __ITEM__ (from the text __TEXT__). Please rework your comment so that it doesn't include the inappropriate items.")
Donate Life
Posted 01 Feb 2005 at 1:38am #
Could you set the string as " cialis" (that is with a space before it) to avoid blocking British, Canadian, etc. spelling?
Posted 01 Feb 2005 at 8:42am #
No. "buycialis.com" has no space.
Posted 01 Feb 2005 at 2:02pm #
Interview with a link spammer | The Register
A thin interview with a link spammer at The Register. Nothing really new or interesting. Link courtesy of NSlog.
...
Posted 02 Feb 2005 at 9:10am #
While this is true, setting up two URL Pattern matches like this will avoid blocking "speci_alise" or "speci_alist" whilst still catching anything beginning or ending with "ci_alis" --
ci_alis\b
\bci_alis
You have to split them up in order for the catch-all-ness to work; if you combine them like this
\bci_alis\b
you'll only catch stuff where "ci_alis" stands alone in the domain name.
(Underscores inserted to avoid MT-Blacklist 😉 )
cl
Posted 02 Feb 2005 at 9:14am #
I'll take my chances and be happy to block buycialisnow.com as well. People are now told what the error is, so they can figure it out.
And as you may note, TypeKey members can post nearly anything.