Wednesday, February 06, 2008

Jonathan Snook:How I built an effective blog comment spam blocker

Arrol ir, hogy egy pontrendszer segitsegevel szuri ki a spameket a comment-ek kozul.

A lenyeg:

For everything in a comment that I like, you get a point. For everything I don’t like, you lose a point (or two, or three). If you get a 1 or higher, you’ve made it on the site as a valid comment. If you get a 0, it’s set for moderation and I’ll take a look at it. If it’s below 0, it’s marked as spam and I’ll never see it (although I check every couple weeks just in case a legitimate comment needs to be unflagged). If it falls below -10, I don’t even bother saving it to the database since it is so obviously spam.

Pontokat meg ezekre ad:

How many links are in the body More than 2 -1 point per link
Less than 2 +2 points
How long is the body More than 20 characters and there’s no links + 2 points
Less than 20 characters -1 point
Number of previous comments from email Approved comments +1 point per
Marked as spam -1 point per
Keyword search Levitra, viagra, casino, etc. -1 point per
URLs that have certain words or characters in them .html, .info, ?, & or free -1 point per
URLs that have certain TLDs .de, .pl, or .cn (sorry guys) -1 point
URL length More than 30 characters -1 point
Body starts with… Interesting, Sorry, Nice or Cool. -10 points
Author name has http:// in it -2 points per
Body used in previous comment
-1 point per
Random character match 5 consonants -1 point per