Spam blacklisting on Foundation wikis

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Spam blacklisting on Foundation wikis

Herby-2
Discovering a backlog of requests on Meta spam blacklist Aphaia & I have
put some work into clearing this and now have it under control (I
hope!).

However in the course of this I have learnt a lot and some (to me)
Foundation level questions arise.

The policy at Meta has been only to blacklist those sites with
persistent cross wiki records of link placement.  However some of the
sites that have been requested for blocking are fairly obviously
undesirable whether they currently are troubling just one wiki or many
(porn being the obvious example but there are plenty more that I think
consensus on undesirability would be found).

So should we be rejecting requests for such sites saying that they
should use their own local lists?  I am aware that preventative blocking
of anything or anyone can be frowned on (personally if I can see trouble
coming I prefer to take action before it arrives rather than clean up
afterwards).  It may be that, in asking en wp for example to block
locally, we are merely missing the opportunity to avoid problems across
wikis in the near future.  Spammers are adept at exploiting any opening
that they can.

There has been a tendency to treat the Meta blacklist as a place of last
resort - I question whether a more thoughtful approach to keeping
Foundation sites clean might not be desirable?

Herby
[[user:Herbythyme]] most places
--
  Herby
  [hidden email]

--
http://www.fastmail.fm - A fast, anti-spam email service.


_______________________________________________
foundation-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/foundation-l
Reply | Threaded
Open this post in threaded view
|

Re: Spam blacklisting on Foundation wikis

Luna-4
Any link blacklist will involve a certain amount of arguing and bureaucracy
-- the meta blacklist affects all languages and all projects (the stakes are
higher), and the relationship between people making requests and people
maintaining the list is different. It's my understanding is that local
blacklists allow us to manage things "in house" and hopefully avoid a lot of
the policy arguments from meta -- the restrictive policy of one wiki is not
a huge concern, anymore, if it doesn't directly affect the other wikis, for
example.

If we keep things local, we keep things simple for the "end user" (editor),
which I guess is the way to go. Treating meta as a "last resort" is probably
a bit much, as you've said, but I don't know that we gain anything by using
the central blacklist all that often.

How often do spammers go after multiple wikis? Multiple languages? How
quickly will we notice, if they do? Those questions seem important to me, in
terms of deciding how often to use/consider the meta blacklist.
Unfortunately, I'm not aware of statistics on the first two. As far as "how
quickly will we notice?" I'm guessing we're counting on people to notice
these things (people heavily active in multiple projects and languages,
especially), which probably means not very quickly. It might be helpful to
have some software try and keep track of cross-wiki spamming -- beyond my
ability, but I bet it'd be helpful, here.

Just a thought. Thanks for bringing it up.
-Luna
_______________________________________________
foundation-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/foundation-l
Reply | Threaded
Open this post in threaded view
|

Re: Spam blacklisting on Foundation wikis

Yann Forget-2
Hello,

Luna a écrit :

> Any link blacklist will involve a certain amount of arguing and bureaucracy
> -- the meta blacklist affects all languages and all projects (the stakes are
> higher), and the relationship between people making requests and people
> maintaining the list is different. It's my understanding is that local
> blacklists allow us to manage things "in house" and hopefully avoid a lot of
> the policy arguments from meta -- the restrictive policy of one wiki is not
> a huge concern, anymore, if it doesn't directly affect the other wikis, for
> example.
>
> If we keep things local, we keep things simple for the "end user" (editor),
> which I guess is the way to go. Treating meta as a "last resort" is probably
> a bit much, as you've said, but I don't know that we gain anything by using
> the central blacklist all that often.
>
> How often do spammers go after multiple wikis? Multiple languages? How
> quickly will we notice, if they do? Those questions seem important to me, in
> terms of deciding how often to use/consider the meta blacklist.
> Unfortunately, I'm not aware of statistics on the first two. As far as "how
> quickly will we notice?" I'm guessing we're counting on people to notice
> these things (people heavily active in multiple projects and languages,
> especially), which probably means not very quickly. It might be helpful to
> have some software try and keep track of cross-wiki spamming -- beyond my
> ability, but I bet it'd be helpful, here.

Most often, they spam several langages and several projects at once,
even within and outside Wikimedia.

I think we need another feature to prevent spamming: the ability to
block the creation of pages with a certern pattern like .*/w/index.php
which are only created by spammers. A central black list of these would
be useful.

> Just a thought. Thanks for bringing it up.
> -Luna

Regards,

Yann
--
http://www.non-violence.org/ | Site collaboratif sur la non-violence
http://www.forget-me.net/ | Alternatives sur le Net
http://fr.wikipedia.org/ | Encyclopédie libre
http://fr.wikisource.org/ | Bibliothèque libre
http://wikilivres.info | Documents libres

_______________________________________________
foundation-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/foundation-l
Reply | Threaded
Open this post in threaded view
|

Re: Spam blacklisting on Foundation wikis

Herby-2
In reply to this post by Luna-4
On Thu, 23 Aug 2007 19:58:32 -0700, "Luna" <[hidden email]> said:

> Any link blacklist will involve a certain amount of arguing and
> bureaucracy
> -- the meta blacklist affects all languages and all projects (the stakes
> are
> higher), and the relationship between people making requests and people
> maintaining the list is different. It's my understanding is that local
> blacklists allow us to manage things "in house" and hopefully avoid a lot
> of
> the policy arguments from meta -- the restrictive policy of one wiki is
> not
> a huge concern, anymore, if it doesn't directly affect the other wikis,
> for
> example.
>
> If we keep things local, we keep things simple for the "end user"
> (editor),
> which I guess is the way to go. Treating meta as a "last resort" is
> probably
> a bit much, as you've said, but I don't know that we gain anything by
> using
> the central blacklist all that often.
>
> How often do spammers go after multiple wikis? Multiple languages? How
> quickly will we notice, if they do? Those questions seem important to me,
> in
> terms of deciding how often to use/consider the meta blacklist.
> Unfortunately, I'm not aware of statistics on the first two. As far as
> "how
> quickly will we notice?" I'm guessing we're counting on people to notice
> these things (people heavily active in multiple projects and languages,
> especially), which probably means not very quickly. It might be helpful
> to
> have some software try and keep track of cross-wiki spamming -- beyond my
> ability, but I bet it'd be helpful, here.

At least a couple of sites I've blocked have spammed 20+ wiki.  There
are some folk on en wp spam project that are very hot on
checking/chasing cross wiki spammers.  I followed one across es, it & fr
wikis a week ago.  There are some great tools out there for cross wiki
contributions
(http://tools.wikimedia.de/~luxo/contributions/contributions.php?lang=en)
and links (eagle's one - can't find the link instantly).

It certainly happens and a few folk are very dedicated to keeping
foundation sites as clean as possible.  I know it isn't a main stream
aspect of the project but keeping porn, gambling, finance links out of
pages seems a worthwhile activity and worth finding a "good" or even
"best" way to deal with it.

Cheers
Herby
--
  Herby
  [hidden email]

--
http://www.fastmail.fm - The professional email service


_______________________________________________
foundation-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/foundation-l