[Open-bio-l] Open Bio Wiki spamming

[Dan Bolser]

This captcha plugin (and rule set) has never let me down:

http://www.mediawiki.org/wiki/Extension:ReCAPTCHA


If you have the 'AskSQL' extension installed [1], you can directly
query the database in order to detect spamming accounts (see below). I
then typically list all the edits by those users and revert them all
before deleting the user [2]. Another strategy is to merge all
spammers into one user and do a mass 'rollback'.

[1] http://www.mediawiki.org/wiki/Asksql
[2] http://www.mediawiki.org/wiki/Extension:User_Merge_and_Delete


There are also other dedicated spam cleanup extensions, but I never
found a 'killer-app' for this job.


Here is a couple of queries that may help:

* Funnily enough, many of the spam accounts use the same email address:

SELECT
  RIGHT(user_email, LENGTH(user_email) - LOCATE("@", user_email)) AS
mail_domain,
  COUNT(*) AS domain_count
FROM
  prefix_user
GROUP BY
  mail_domain
HAVING
  domain_count > 1
;


* Your top editors may well be spammers...

SELECT
  *
FROM (
  SELECT
    rev_user_text, COUNT(*) AS EDITS
  FROM
    prefix_revision
  GROUP BY
    rev_user_text
) AS inner_table
ORDER BY
  EDITS DESC
LIMIT
  20
;


HTH,
Dan.


P.S. Since we're talking spam ;-)

NETTAB 2010 workshop focused on: Biological Wikis!
November 29 - December 1, 2010, Naples, Italy
http://www.nettab.org/2010/


On 7 August 2010 07:07, Dave Messina <David.Messina at sbc.su.se> wrote:
> Is it possible to tell whether the spam account was created by machine and not human?
>
> If a human, better captcha might not help us.
>
>
> Dave
>
>
> _______________________________________________
> Open-Bio-l mailing list
> Open-Bio-l at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/open-bio-l
>