On Tue, Mar 1, 2011 at 23:49, Alexander Lamaison <[email protected]> wrote: > On 1 March 2011 22:06, Peter Stuge <[email protected]> wrote: >> Alexander Lamaison wrote: >>> > The spam thing is pretty much impossible to solve perfectly in an >>> > automated way.
This is unfortunately true. Spamming has changed. A (normally) very low-traffic wiki I visit and use was recently totally overflowed with the most useless, pointless spam over a period of a month and a half, until the admin was forced to totally disable account creation (now it's a manual process - you mail him, talk to him, he learns who you are, and he adds you). The spammers were obviously human, not bots (if you deleted the content of a newly created spam page they would notice and revert your fix), captchas and email authentication tickets were duly answered, and they created hundreds and hundreds of accounts at the rate the old ones were blocked. They added mostly job listing pages, but what was baffling was that every page was a new page, thus there were no links there from anywhere inside the wiki (you could mass delete them just by going to special:orphaned and look for them). So it looked utterly pointless. If they just wanted free storage which could be linked to from elsewhere there are free and easy legal ways to do that. Totally baffling. They started with A and went through the alphabet (Alabama jobs, Christian Teacher jobs, and so on), and started over when they reached the end of their spam job listings. To me it looked like a couple of guys were paid to sit 16 hours a day doing this. And they used a network of proxies so that IPs were always different. I can't see any well-functioning automated way to handle spam done manually like that. -Tor (p.s. What I really want is a global task force with resources to physically hunt down every spammer everywhere and do bad things to them.) _______________________________________________ libssh2-devel http://cool.haxx.se/cgi-bin/mailman/listinfo/libssh2-devel
