Forum Moderators: phranque
I've tried a combination of different methods to try and stop this: watching for more than x positive feedbacks for a particular user from the same IP address and/or user agent, insisting that new members can only leave feedback after x days, personally reviewing all feedback and related activity to see if I think it's genuine... these methods are either inpractical, give too many false positives or are too restrictive on the user.
anyone have any other ideas?
For example: I could register at a site as myname@mydomain.net, mydog@mydomain.net, mycat@mydomain.net, etc. and I get all those e-mails and can respond to each as a separate account. (But I'm a webmaster and, like all webmasters, I am unintentionally the bane of other webmasters everywhere -- EXCEPT when I am on webmasterworld.) Never design a site to deal with webmasters... I bet Brett would agree on that. :)
A lot of the fakers don't provide real addresses, so we require email verification in two ways. First, they have to input their email address twice. A server-side script tests to be sure the addresses agree. This is mostly to be sure that they haven't made a typo.
Then, the site automatically sends them an email to the address they gave. There is a link in the email that they have to click on in order to have the registration "activated" on the server. The new registrants don't become fully active members until they click on that link. That means that they can't log in until they're "active".
You'd be amazed by the high percentage of people who never complete the process. Some probably don't get it, even though it is explained on the site and in the email. Some are clearly fakers. After a period of time the non-actives are removed from the system.
Of course, this won't stop people from registering twice with working addresses.
If Amazon can't stop shilling [dictionary.reference.com] than you might be out of luck!