1

Hi there!
So I've noticed an increase in Spam posts, even with the occasional Spam Bot stop that we get. So I'm wondering if there is a way to flag posts for moderation queue when the system detects words or phrases associated with commercial adverts. This way, we won't ever have to see the posts, because the moderators will see it first, and then they would remove it automatically.
If something like that has been thought of, or if it's being worked on, I would like to know about it.
Thanks!

Thumbs up

2

Even if somebody was building something like this, it would be a futile effort because it would be too easy for the spammers to defeat.

Thumbs up

3

How would the spammers defeat this? All we need is something similar to what Facebook uses to detect web sites that may not be secure. For instance, if I sent a link on Facebook to my friend, and Facebook blocked it because it could be unsecure, perhaps something like that could be implied as well.

Thumbs up

4

No bit.ly links in your first post. Oh, hey, and maybe newbies have to go through moderation anyway? It wouldn't be practical on a very large forum, and I'm not sure how much work the mods are having to deal with already, but it seems like newbie posts getting put in moderation until they are approved might help? Especially if there are links. Or an overabundance of spelling/grammar weirdness, but that'd be harder. Ah, but I don't think many first posts have much reason to go outside the first 255 Unicode characters, and we have had spambots that don't even pretend to speak English, so maybe anyone not explicitly approved gets put in moderation if they post links, outside the character set the mods can read, or within 10 seconds of registering?

Actually, is there a cooldown period between posts already? If a human needs to post more than once within 10 seconds, other than maybe editing their own post, I'd expect that to be very unusual indeed.

Some of my games
Keep up to date by following @Jeqofire on twitter!
Ear Ninja?

Thumbs up

5

Back when I administered the Armadillo user's forum, moderating a user's first five posts was the approach we eventually took. While the moderators and I still had to deal with them, it did keep their trash off the forums.

I discovered that I had to make one tiny little change to the forum software which the vendor started including in their official releases. That was to only count approved or visible messeges in the user's count of posted messeges. Otherwise a spam bot could defeat us by posting more than five messeges.

Thumbs up

6

We already have several reputation-based countermeasures, but inevitably these will be bypassed. And we also have some tools for dealing with potential new floods of spam from given IP addresses or forum posters when they are identified, at the time they are identified.

But basically, we're just as aware of spam as you are. That is, either we see the spam ourselves, or somebody reports it. Apart from, as suggested, a pre-flight moderation period, I really can't imagine what we would do to improve this situation aside from detecting particular features of posts, and all the ones I can think of or that have been suggested over the years are either controversial or likely to be trivially defeated. It is also the case, and this might not be apparent to most, that many spammers actually are real human beings--that is, they actually build up a reputation manually, before flooding the place with spam. This is simply not something we can do anything about because it's clearly intentional.

We'll keep on the lookout for options for identifying spammers before too many people notice, but inevitably some of it is going to slip through the countermeasures we already have. There are discussions on the moderator mailing list as well about improving the software so we can respond more rapidly when the spam is identified. In the mean time please keep reporting the spam you do find. I can assure you that we take great pleasure in nuking it. smile

Just myself, as usual.

Thumbs up

7

@green gables fan

What facebook uses is probably a database of known IP sources of spam and where spam bots are known to operate from. Which means somebody has to maintain and update it. It's more than likely a service that facebook pays a subscription to so they can use it.

Thumbs up