1

Possible Duplicate:
How can I prevent spam on sites which I control?

Are there any recommended admin tools or interfaces for reviewing and moderating spam on a community-driven UGC site?

Improve this question
4
  • 2
    This isn't the right forum for this type of question.
    – wheaties
    Commented Sep 30, 2010 at 14:37
  • 1
    spam what? Spam email? Spam blog posts? Spam tweets? There is a great algorithm for determining if something may or may not be spam, but how useful or practical it is depends on your situation. en.wikipedia.org/wiki/Bayesian_spam_filtering
    – steven_desu
    Commented Sep 30, 2010 at 14:38
  • Sorry about that, I clarified the question. I also found this, which somewhat answers the question: meta.stackexchange.com/questions/4565/…
    – Albert Wang
    Commented Sep 30, 2010 at 14:39
  • Is the "c" tag really relevant?
    – Lèse majesté
    Commented Feb 21, 2011 at 1:54

4 Answers 4

Reset to default
2

There aren't "best practices" for that. To stop spam you need to be creative. Don't use the traditional type of captchas which ask the user to enter some text from an image into a text box. As an owner of a big forum hosting company I can guarantee these Captchas do not stop spambots, not even the hyped ReCaptcha.

Think of some security question anyone would know how to answer. For example: "Which number is larger: 1 or 328?". For a number of reasons this is a fairly effective way.

Akismet is also nice but that is if you don't depend too much on user input because it often flags legit comments as spam comments. So it's not useful for forums, only for blogs and similar websites where you can afford to lose an user or 10.

Improve this answer
7
  • 1
    A CAPTCHA is any type of automated test to distinguish bots from humans, and they do work. But no CAPTCHA system is 100% foolproof, including the one you proposed--which only works on obscure sites which spammers don't pay much attention to. Otherwise it would be very easy to write a bot that can pick out the larger of two numbers or solve an algebra equation.
    – Lèse majesté
    Commented Feb 21, 2011 at 16:02
  • 1
    The problem is that any test that a computer can generate automatically, a computer can also solve automatically. So the really effective CAPTCHA tests either have to be manually created or they rely on obscurity. That's just not practical for large sites. But you're right that for small sites, you can easily stump bots by coming up with a creative CAPTCHA test that isn't in wide use. Industry-specific questions are another good example of this.
    – Lèse majesté
    Commented Feb 21, 2011 at 16:06
  • Vergil what you listed as a 'better example' is actually a CAPTCHA. CAPTCHA is : Completely Automated Public Turing test to tell Computers and Humans Apart.
    – XOPJ
    Commented Feb 21, 2011 at 16:29
  • 1
    That's a variation of a CAPTCHA solution. Can you please reword your answer so it clearly states you feel that variation is flawed. I agree it is, but your answer can be misleading to future people that find this Q/A. As old variation techniques are retired new ones come out, such as "click the cat" or "solve this math problem" or "which makes you feel happy" type questions. The biggest problem with non-text based CAPTCHA's is language. Typing in matching characters is easier to transcend language then which is fluffy:"kitty, cactus, fire".
    – XOPJ
    Commented Feb 21, 2011 at 16:38
  • 1
    Okay, I edited my answer to make everyone happy. :)
    – Vergil Penkov
    Commented Feb 21, 2011 at 17:49
1

I've always had luck with SpamAssassin. http://spamassassin.apache.org/

Improve this answer
1

While perhaps a little dated, this is a good starting point if you're trying to think about developing something: A Plan for Spam

Improve this answer
0

Usually moderation tools are built into the application/CMS you're using. If you're talking about automated spam-detection tools, then Akismet has a pretty good reputation for blocking spam on blog comments (you can use their API for any type of content though). And then there're also CAPTCHAs and bot-traps, which can also be very effective when employed strategically.

User policies are another good way of preventing spam. For instance, peer moderation works quite well on large sites. StackExchange's policy of requiring a certain amount of reputation before allowing links/comments is also very effective.

When all else fails, hire some paid moderators to patrol the site and delete spam posts and ban spammers.

Edit:
Virgil brings up a good point about the ineffectiveness of the popular text-recognition CAPTCHA technique. I can't find any good stats on the latest ReCAPTCHA tests, but most advanced bots have good enough OCR to defeat them within a few attempts. Luckily, the smaller the site you operate, the less effort spammers will make trying to defeat your CAPTCHAs. So for many sites, a well-designed text-recognition CAPTCHA paired with a bot-trap field is enough to deter most spammers. But if you operate a larger site, you may need to get more creative or use an image-recognition CAPTCHA like Microsoft offers.

Improve this answer