Re: [website-dev] One more step in the Spam battle on the OOoWiki

On 12/10/2010 09:08 AM, Andrea Pescetti wrote:
> Clayton wrote:
>> With some (a lot actually) help from the MediaWiki mailing list guys,
>> I've added some code to watch for two of the common throwaway email
>> services.  This can be expanded if necessary by adding to the reg.ex. in
>> the function. This is now documented here:
>> http://wiki.services.openoffice.org/wiki/Help:Block_email
>
> Honestly, blocking humans that deliberately want to spam seems a very
> difficult task to me. For one, http://mailinator.com/ shows on its home
> page a list of several alternate equivalent domains, and the regular
> expression catches only two of those.
>
> So either you try to extend the regular expression to catch more of the
> alternatives (but then it will need updating from time to time) or you
> accept to block only a part and agree that a deliberate spammer will
> manage to spam the OOo wiki (which seems unavoidable to me).

You're right, mailinator does have several alternate domains, and the 
spammers can still at this point use those domains.  We can easily add 
them to the function if they are used.  I could have limited all domains 
that mailinator provides, but initially, I am just trying the two most 
common to see how it works.

There are several people monitoring the Wiki for spam. Whenever I spot a 
spammer, or notice that one of the other moderators blocked a spammer, I 
take a quick look at what edits they performed and what email they used.

The edits are usually to add links to websites.  I add the websites to 
the blacklist so they can no longer add that specific website.  This has 
really reduced repeat spam.

The email addresses being used (up to today) have been almost 
exclusively mailinator.  Occasionally they use a legit webmail address 
at hotmail or gmail, but it's rare.

Spam and vandalism is unavoidable, but each step taken has progressively 
reduced the spam and vandalism.  If you look at how much we get now vs 
even just a year ago, there is a noticeable difference (much less now).

C.