Automated account reputation, spam removal and posting rights
Posted Nov 20, 2003 12:02 UTC (Thu) by copsewood
Parent article: On comment abuse
I suggest LWN combine allowing almost unrestricted commenting by subscribers with the need for non-subscribers to do a Turing test (e.g. recognise a number in a graphic) and be filtered based on account reputation. Subscribers could be enabled to rate a limited number of all posts per day on a scale between 1=spam, 2=troll and 10=profound. 3 hits on the spam button should be enough to have the post automatically removed and reported to the editor - subscribers who in the opinion of the editors use ratings very unfairly, post spams or more than one or 2 trolls could have their ratings and/or posts silently ignored.
Perhaps an easy way for a non-subscribing account to build an acceptable reputation to be allowed occasional posts is simply for it to exist and be used regularly for a period of time. If someone is paying attention, the chances are that what they have to say is more likely to be worth saying, than if they have just created an account, or not used it regularly for any length of time before attempting to post. I think this, together with letters to the editor, could be all that is needed in the short to medium term, but if this doesn't work well enough, e.g. because trollers are using automated means to create accounts and build reputation scores on these based on usage then either more regular use of Turing tests or some kind of IP firewalling might be needed.
to post comments)