Volca on 30/10/2008 at 11:40
Maybe this is a bad idea, but I'll say it anyway:
Would it be feasible to check first post of each subscriber?
Maybe it would be better to check the first/second posting instead of the account creation - the registration is often done by human or with a human assistance (captchas linked to warez/erotic pages so they get the captcha filled in by human for example).
Accounts posting spam could be removed automatically (maybe vbulletin has an option to use some bayesian filter or spam assassin itself? That would mean it could get automatized).
It's a pity so much work has to be done in order to keep the forums free of this.
pavlovscat on 30/10/2008 at 18:28
And a big thanks to David for all his hard work keeping all that crap off the forums! :thumb:
And of course everyone else!
R Soul on 31/10/2008 at 14:46
I think a more reliable means of captcha would be to ask the user a question about the image. Say, some letters are upside down and the user had to say how many are like that.
Or what about two questions: How many letters are upside down, how many are there in total, and the user must wait for at least one minute after answering the first question before they can answer the second.
I imagine these spam bots are aimed at lots of sites, so if you increase the time it takes for a human to help them along, would that not frustrate their efforts?
baeuchlein on 1/11/2008 at 12:47
Quote Posted by pavlovscat
And a big thanks to David for all his hard work keeping all that crap off the forums! :thumb:
A big
Thank You to
all the people here (including, for example, the moderators) doing this and other "behind the scenes" work to keep the forums up and running!
jtr7 on 7/11/2008 at 02:15
Is it just me or is TTLG getting hit with another big wave, right now?:mad:
Yes. Much gratitude. Maybe we should chip in for a case of fine brew for our Administrators extraordinaire.
Kolya on 13/11/2008 at 03:33
These spammers have an easy life because our sole defense are a few moderators and they are many, with endless supplies.
The "Report this post" link is a way to tap the potential power of the community. A pretty weak tapping that is.
It's like a dark knight rides into your village and all you can do is point at him and hope one of the white knights comes by. You don't even see if others are pointing too. It's unsatisfying.
When I click the "Mark as spam" link on YouTube that comment gets hidden right away for me at least. When I rate it down other see that as well.
But when I click "Report this post" I have to fill in some stupid message box and then squat shit happens that moment. It should at least hide that post for me and preferably it should get a marked-as-spam-flag that's visible for other users.
This is no criticism whatsoever, just some thoughts on how it could work. (My first thoughts circled around running concerted DoS attacks on the spammer's site.) In any case, tap that community power. I think that's the only way.
Gingerbread Man on 13/11/2008 at 22:28
As long as it didn't get abused, I think a member-level "Report as spam" link that could quarantine the flagged post and temporarily ban its author (pending admin deletion) if it logged a certain number of reports -- especially if it could be sensitive to a large number of reports in a short period -- would be a very useful thing.
It'd solve 80% of it if we just banned Chinese and Russian IPs outright, tbh.
(edit: Okay, I just did another load of validations. It'd solve more like 95% of it if we banned .cn and .ru)
Scots Taffer on 13/11/2008 at 23:32
For the remaining 5% I'd like to opt in for moderator control of spammers :cool:
Nameless Voice on 14/11/2008 at 00:07
Quote Posted by Gingerbread Man
(edit: Okay, I just did another load of validations. It'd solve more like 95% of it if we banned .cn and .ru)
Would it be possible to configure the forum to only require manual verification from people with .cn and .ru? Sounds like that would save you a lot of work.
David on 14/11/2008 at 07:14
Nope, a huge chunk of them are GMail.
