Unbanning the Unjustly Banned!

I am getting some emails that the banning software is acting up again and banning people. Please confirm you're banned (the system will give you a specific message stating that your IP address -- a series of twelve or so numbers -- is banned).

And then send that IP to me. And I will get you unbanned.

I just wrote to Kemp:

I've tried to explain this: We have an automated banning system which
bans spam. Sometimes a spammer fakes your IP. (They fake a lot of
IPs.) And then it gets banned.

And sometimes you just accidentally set off the software's tripwires
by using a couple of words that spammers use. Like, maybe, it's
learned "Lace Wigs" is spam so if you do a joke post about Lace Wigs,
now you're banned, because the software obviously doesn't understand
irony or parody.

I will contact Pixy immediately about unbanning the IP. He's on New
Zealand time, though, so this might take a day or two... our emails
tend to cross without being read.

That last part is important, and a reason that some people who I previously said I'd get unbanned remain banned. Pixy's on a whole different time-cycle, and further, he's got a real job. He does this hosting stuff as a favor, and because he's a conservative who wants to make a difference.

But what happens, sometimes, is that I write him and he misses the message. Happens to me too -- I get a lot of emails.

And then I assume the banned IP has been unbanned... but sometimes it hasn't been.

So... what I'm going to do is do a big email of a bunch of banned IPs, and make sure he sees them. Like, if I don't get a reply, I'll write back in a day.

So please email me your banned IP addresses with the subject line, in caps, UNBAN and then your IP address.

Should take a day. If he misses the email, two days. If he misses that one, three days... but we will get you all unbanned.

Oh: I should say I am making that up about the "Lace Wigs" example. I don't think that's really a word-pair that will get you banned. I'm using that as a for-instance. The system has some AI script it uses to determine if a comment is spam. Words used, rapidity of comments being posted after each other, stuff like that.

The system screens out a lot of spam -- you can see that it still permits an awful lot of spam, too. Without the system, we'd have ten times as much spam.

But it frequently bans real commenters too.

Trust me -- when I ban people, I tend to make an announcement. I don't do stealth-bans.

posted by Ace at 01:45 PM

| Access Comments