[libreoffice-website] Ask site: Canning our SPAM problem

Hi everyone,

We've been dealing with a bunch of SPAM on the Ask site over the last
couple of weeks. We recently enabled moderation on user content in an
attempt to stem the vandalism of the front page, and I've spent the
last couple of days investigating some possibilities for the site. As
you might expect, I haven't found any silver bullets, but I do have a
few suggestions that can help us manage the problem.

Based on my research on the Askbot.org site, a number of SEO-driven
spammers have taken to posting on Askbot sites. It's not clear what
percentage of these postings are run by scripts, but it's my
understanding that some of the spammers are using some form of
automation. On the AskLibo site, we're seeing not only SPAM in
questions, but we're seeing SPAM in _follow-up_ answers as well. These
guys are organized, and probably aren't just going to go away.

The Askbot software includes a few different mechanisms for dealing
with spam and malicious users:

- Content Deletion
- User Bans/Blocks
- Content Moderation
- Registration Limitations
- Akismet SPAM protection/detection

Content Deletion and User Bans/Blocks are pretty straightforward.
Mods, Admins, and some users with high karma may delete content from
the site. Users with less karma may flag content. All users spamming
the site are banned (it's been pretty cut-n-dry up to this point).

Content Moderation entails all user content going into a queue that is
reviewed by the mods and then accepted for the site or deleted. This
is what we have enabled right now.

Since we've enabled Content Moderation, the spam on the site has
basically disappeared [1]. We've approved a lot of content by
legitimate users, and we've banned and deleted a number of
spammers/spam content.

> From the moderator side, Manfred (@manj_k) and I both agree that
enabling moderation has given us more work, since we want to make sure
that content does not sit for long in the moderation queue. That being
said, before moderation was enabled, SPAM would sometimes sit on the
site for hours at a time, and I've heard multiple LO
users/contributors refer to the site as a "SPAM pit" or "totally
messed up". We've also run into some hiccups with the moderation
interface which prevented us from actually being able to approve any
questions.

> From the user side, we've only gotten feedback from a few Ask users.
Reception was mixed, with an overall feeling that moderation is an
acceptable short-term solution, but needs a better long-term fix. One
of the users suggested... (a perfect segue)

...Registration Limitations

Other sites have used registration limitations to reduce/eliminate
spam on their Askbot site. One mechanism suggested to us was to
require our users to have an email address from a "paid" provider.
That is, no gmail, yahoo, hotmail, etc... accounts would be allowed
(at least not during registration).

The suggestion is clever and powerful, and could probably serve as an
excellent filter on spam, but my concern with such a proposal is that
many legitimate users rely on a free provider for their email, and
such an hard rule might discourage many of our meekest users from
participating on the site and getting help from us.

Next up in our quiver of features is Akismet SPAM
protection/detection. I don't know too much about this product, but
it's basically an external SPAM-filtering service. People rave about
it on this wordpress page[2], and I've seen it in action on a
Wordpress hosted blog. To quote Ron Popeil, you "set it and forget
it". There are two possible issues with Akismet, (1) it ain't Free,
and (2) it ain't free.

(1) As a hosted, commercial solution, Akismet is running some
proprietary software stack. So no Free/Libre/Open Source Software love
here.

(2) Akismet is free for (some) personal use, but costs something more
in all other cases[3]. That might be $50/month, $100/month, or they
might just be nice and give us some kind of discount. It's pretty
nebulous here.

I'm not sure whether (1) or (2) would preclude Akismet being used in
the LO infrastructure, but I figured I'd include the software/service
in the list and let someone else tell me the rules :-)

---

That mostly sums-up what's present in the Askbot software *right now*.
There are some other options that have been proposed for future
development in Askbot including IP-based tools (Blocking by IP, User
identification by IP, etc..), karma-based moderation, and group-based
moderation. Some of these ideas are discussed on the Askbot site for
Askbot[4].

I've asked a couple of questions on the Askbot.org site[5] and have
received very prompt replies from the creator of the software, Evgeny
Fadeev. One proposal I made was to allow mods to tweak the application
of Content Moderation so that it only applies to a certain subgroup of
users, for example:

 * All users with karma < MIN_KARMA_TO_SKIP_MODERATION
 * All users who are not in a group flagged to SKIP_MODERATION, or
 * All users whose account is younger than MIN_AGE_TO_SKIP_MODERATION

I've had a couple of chats with Alex Werner, and he believed that we
should be able to put the bulk of our users into a 'trusted' group so
that they can skip moderation. He also insinuated that karma-based
moderation might be available in a newer version of the Askbot
software, and that we might be able to upgrade in the near future.

As I said in the first paragraph, there's no silver bullet here...at
least not *yet*. We can try out a new version of Askbot, test Akismet,
or maybe make some local tweaks to our software. Most importantly,
we've got the 4.0 release coming up very soon, and I'd really like the
site to be ready to receive a bunch of new users and help answer their
questions swiftly and professionally. For now, given that we don't
have any way to bypass Content Moderation for our existing users and
given that we've had some system instability with the moderation
interface, I suggest that we disable Content Moderation. I'll try to
check the site regularly, especially next week, and encourage all our
users to flag SPAM early and often.


Cheers,
-- Robinson 'qubit' Tryon
LibreOffice Ask site moderator


[1] Side note: Due to some weirdness/bugs in the system, we've
actually had to push content *through* the approval system before we
can delete it, but it's only live on the site for a few seconds.
Hopefully the latest version of askbot will fix that.
[2] https://wordpress.org/extend/plugins/akismet/
[3] https://akismet.com/signup/
[4] http://askbot.org/en/questions/
[5] 
http://askbot.org/en/question/9882/is-it-possible-to-moderate-user-content-based-on-karma-or-other-factors/

-- 
Unsubscribe instructions: E-mail to website+help@global.libreoffice.org
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://listarchives.libreoffice.org/global/website/
All messages sent to this list will be publicly archived and cannot be deleted