Google Censors Socialtext
Jonas Luster blogs in depth about how Google is censoring Socialtext under the label of Badware.
All in good fun, and we have, of course, filed a proper appeal. And it isn't that StopBadware.org is a bad initiative. My personal problem is I find a lot of truth in the above headline. There is no report about Socialtext in StopBadware's database. Yet, Google says we are in it.
Sounds like someone isn't fulfilling their mission, censoring an erstwhile competitor and while they may not be evil, they are acting badly.
Elsewhere: the road to hell is paved with good intentions
Linking is not a crime
While the above is all fun and good, we think we were censored because one of our public wikis enabled someone to attach and HTML file that contained a link to a site that has badware.
So, let's see if we can get Typepad censored by making the same link. (UPDATE: I removed the link because I don't really want to get Typepad in trouble and think my point is made)

That link, plus whatever script on your page shows popup preview windows of links on over, results in badness.
FYI.
Posted by: dbt | January 05, 2007 at 12:53 PM
Chicken!
Seriously, if your hypothesis is true then that would hamper public discussions in the security community. But then we've typically 'shared' only zipped copies of attack examples, not links.
Perhaps a bridging website for dangerous links is needed, something that warns the visitors clearly and recommends safe viewing conditions (i.e. vm browser session) before proceeding. It can also hide bad links from badware link detection bots.
Posted by: Don Park | January 05, 2007 at 02:52 PM
I asked someone about this. There was a lot of bad/malware looking content under the /stoss/ subdirectories. Urls that started like
http://www.socialtext.net/stoss/index.cgi/emma-watson-topless.html?action=attachments_download;page_name=kinose... http://www.socialtext.net/stoss/index.cgi/gallerie-lesbiche.html?action=attachments_dow...
http://www.socialtext.net/stoss/index.cgi/mariah-carey-nude.html?action=attachments_downloa...
http://www.eu.socialtext.net/page/stoss/using_strut_to_import_files_into_a_socialtext_workspace/attachments/20061127141711-7-18582/film-porno-gratis.ht... and so on (I pruned the urls because clicking on the full urls can be pretty dangerous).
So I think it was accurate that there were some pretty bad urls on the site that users might to get a warning about.
Posted by: Matt Cutts | January 06, 2007 at 01:15 PM
Thanks, Matt, that's what we figured.
But as a user, I don't appreciate Google interstitling the web for me with potential false positives.
Those user/spammer contributed attachments were removed through regular wiki gardening. Sometimes wikis get goatsed too.
As a wiki provider, if I'm censored there should be a clear method to figure out why, and of course a recourse.
I'm lucky enough to be able to blog and get attention to problems (it is fixed for now). I'm not lucky enough to be able to ask around like you so effectively.
There isn't going to be a perfect solution to this general problem. One way to make it better might be for Google's interstitle warnings to link to the specific badware report the fiter is based on, and have that report offer up recourse.
Posted by: Ross Mayfield | January 06, 2007 at 04:24 PM
In reference to your comment: "even without RSS attention is certainly managable at current scale and noise (like the complexity without the courtesy of letting others know why you are occupying the namespace and others attention, although all intentions probably good). Sharing namespace leads to social discovery and network effects for learning, something I think we all want."
contributed by Ross Mayfield on Jan 10 10:32pm
I'm not as conversant as you with collaborative sites such as socialtext offers, and I don't quite understand the statements you've made here at all regarding 'noise', and ...'courtesy and notifieing people as to why I'm occupieing namespace?????
Posted by: Terry York | January 11, 2007 at 11:18 AM
Meanwhile in Denmark: A blogger received a $970 invoice from a map company for using a deep link to a map displaying the location of his home. They claim that linking to anything other than their home page is a violation of their terms of use and charged the poor guy an amount equivalent to their standalone CDROM version of their maps.
Posted by: Lars Plougmann | January 30, 2007 at 06:45 AM