Can’t some sort of “cop-bot” be automated on the forum’s side so as to deal with many of these cases on it’s own?
There's
MindlessElectron, but I'm not sure what it does:
That account is indeed a bot account, which is owned by me. It used to be called "Mitchell-Bot", but I didn't like that name so I've asked theymos to change it to "MindlessElectron". It will be used to combat spam from a dedicated account instead of my own (and thus risking my personal account).
I'll update the account information later on today.
I mean, when one reports a case as spam it could end up (prior moderator/admin clearance) in a table on the forum’s database, that is then used to filter out posts either on-the-fly
I can think of several ways to sabotage this:
-A spammer can post the same post as a real user, hoping that user gets banned too
-A spammer can easily randomize his post enough, so that a bot won't detect it
-A bad guy can trick an innocent user to post a homograph attack. The user won't even know he did it.
- Auto-banning account (temporary/permanent) with x posts deleted due to spam in a certain period of time (alternatively, add a poop badge on the user profile for a certain amount of time).
This could work, but can be abused too by searching through someone's post history to get x posts banned in a short time.
It shouldn’t be that hard to start off with the trivial cases of exact copies that we see every day, and move on from there to text variants with similarities above a certain degree.
True. But spammers are already adjusting their "strategy" to prevent manual detection. Bots will be fooled even easier.
