On Mon May 04, 2009 at 12:39:06 -0400, Mark Smith wrote:

> We have no automatic spam detection.  Right now it's a purely manual
> process where users delete comments/posts and have the ability to flag
> them as spam.  These posts show up in a spam management queue that the
> support volunteers can process.

  OK, that matches pretty well with what I expected.

> I'm familiar with your service!  I believe it was Denny that pointed
> it out to me some time back?  Not sure.

  That sounds likely, it was Denny that introduced me to the fork/site a
 while ago.

> We're really interested in fighting spam, but haven't put much thought
> or time into the problem.  If you're interested in spam in the context
> of something like Dreamwidth, I'd love to chat.  You can find me on
> AIM as 'xb95', or IRC (irc.dreamwidth.org #dw) also as xb95.

  I'm definitely interested in seeing how that could be done.  I would
 imagine a good starting point is seeing the comments/entries which have
 been reported, and verified, as spam.  From that corpus we can look to
 see how liable automated detection is going to be effective.

  As I learned when I started doing comment detection it is a tricky
 problem, and far far harder than email spam.  (Because you don't get
 clues such as "HELO blah", and the similar meta-data.  You literally
 receive a piece of text and an IP address.

> This is also the kind of discussion we should have on DW itself --
> probably in the dw_dev community.  If you don't have an invite code,
> ping me off-list and I'll get you one!

  I received one today, username 'skx'.  I'll join the dw_dev community
 this evening when I'm back online properly.

Steve
--
http://www.steve.org.uk/
_______________________________________________
dw-discuss mailing list
[email protected]
http://lists.dwscoalition.org/cgi-bin/mailman/listinfo/dw-discuss

Reply via email to