views:

60

answers:

3

i m thinking of making a website..and i want to keep a track on the sent messages! what if some human and not a computer is sending messages that are nt at all subject related..any way to stop that! [apart from the public cooperation]

as in how does stackoverflow keeps track of spams.. how it makes sure that the message sent is totally to the point and subject related!

A: 

CAPTCHA. It's implementation depends on what software you're using. reCAPTCHA is often used as it's easy to setup and use.

Lekensteyn
CAPTCHAs are usability killers...
klez
+1  A: 

Every time a user submits something through your website you should check for spam. You can be sure that as soon as your web site is on the web, spam robot will crawl trough it and submit every form they see.

Now, there are various ways to counter their attack. A very simple way is to add captcha to your form. But captcha aren't that effective and user tend to hate them. Another way is to let them submit, but then an administrator must approve their submission. Other site uses email to make sure user are authentic (forum registration for example). Other site blocks IPs that submit too many forms in x amount of time.

As you can see there is a lot of different techniques (and I haven't list them all). You need to choose the one that would work best for you.

Thierry-Dimitri Roy
OK..what if some human and not a computer is sending messages that are nt at all subject related..any way to stop that! [apart from the public cooperation]
hopefulLLl
@Thierry-Dimitri Roy- how can i make sure that the message that a user is sending is nt abusive and is totally subject oriented..ny way to solve that?
hopefulLLl
@hopefulLLl: have an admin mark one message as "SPAM". Then any other submission from the same user (IP or email) will be flagged as spam. Even better would be to let the users of you site do it for you. Like StackOverflow's "Flag" command.
Thierry-Dimitri Roy
@Thierry-Dimitri Roy -what if the message that is being is sent is to be delivered to a high profile person..since he wont take this abusive message and me being the owner of website cant even let him go through it..isnt there any other way..because what i want is that message shall be private too!so cant really let the users see it!
hopefulLLl
@hopefulLLl then you solution would be to build a system that "learn" to detect message that should not go through. Something similar is done for email filtering. Check "Bayesian spam filtering" (http://en.wikipedia.org/wiki/Bayesian_spam_filtering)
Thierry-Dimitri Roy
+1  A: 

One thing I have done is use a time interval for valid messages. That is, start a timer when the form loads, and have a set determined interval (based on the length of the form) between when a response would be considered valid, and when it wouldn't. If it falls outside of those bounds, then prompt with a CAPTCHA as others have suggested. It's just an extra layer in the honeypot.

jer