Re: SEO spam

[Simon Willison <cs1spw@xxxxxxxxxx>]

Asbjørn Ulsberg wrote:

> I've just gone through ~100 pages on the Wiki and removed the SEO
> spam were inserted by "px2cv.gv.shawcable.net". I hope it's possible
> to track this bastard down, and block his IP or something.
>
> I know it's possible to revert versions, but I didn't know how to do
> it, and I wanted the changes to be done fast, so the SEO bastards didn't
> get what they spammed the Wiki for; hits on their webpages. I hope no
> search engines have picked up the SEO addresses from the Wiki during
> the time the links were there.
>
> To view a page that was spammed, here's the AuthorElementDiscussion
> page of 05.10.2003 08:32:00:
>
> 
http://intertwingly.net/wiki/pie/AuthorElementDiscussion?action=recall&date=
> 1065335520
>
> At the top, right beneath the table of contents, there are three links
> pointing to SEO stuff. I've removed all those on every page I could find.


I just removed another one - very carefully inserted half way through a 
sentence towards the bottom of a page. This guy's smart. As far as I can 
tell, the only way of defending against this is to have a small army of 
people watching the RecentChanges page.

Here's a technical solution (adapted from something I'm considering for 
my blog): How about a mailing list which is sent a diff every time a 
page is edited. At the bottom of each email is a magic link which, when 
clicked, reverts that page back to its previous state (and maybe bans 
the IP address that made the change). If the page has already been 
reverted by someone else, nothing happens. If the page has been edited 
by a legitimate Wiki use since the spam was added, the person clickign 
the link gets a message to that effect and the edit form for the page 
instead so they can manually remove the spam.

Couple this system with a clear message on the edit form in the live 
Wiki along the lines of "Any commercial marketing messages will be 
promptly removed" and hopefully it will act as enough of a deterrent to 
get the problem under control.

Another possibility could be a domain blacklist, like the one I've been 
running on my blog: http://simon.incutio.com/blacklist.txt I@m beginning 
to doubt the effectiveness of this approach simply because spammers have 
so many domains they are trying to approach - I've blacklisted nearly a 
hundred now and that's just on my personal site!

Cheers,

Simon Willison
http://simon.incutio.com/