[cgi-wiki-dev] Important development targets; spam defenses

[Tim Sweetman]

*Unlurks*

Greetings.

Earle Martin wrote:

>What would you nominate?
>
>I'll start the ball rolling: I think we need a banned-content plugin.
>Regular expressions (e.g. "cheap-viagra-levitra.com") could be added to a
>list.
>
> This list would be checked against edit text before the node is
>written. If a match is found, an "Edit Denied" error is shown. Optionally:
>the IP address is also added to a banned IPs list that are not permitted to
>edit. It would be fairly trivial to set up something that gets the big list
>of http://www.communitywiki.org/cw/BannedContent on a nightly basis.
>  
>
>The need for this is clear, IMHO, given that OpenGuides sites are being
>subjected to a barrage of spam at the moment which comes from a range of
>different IPs every time - simple IP blocking will not defend us against
>this.
>
>Thoughts? 
>
Banning IP addresses may constitute a wild goose chase, since 
blocking/monitoring them is the obvious approach, and spammers will 
probably get around it by all the obvious means (botnets, open proxies, etc)

>Other suggestions?
>  
>
Use wiki as "ham" corpus, maintain "spam" corpus, use Bayesian analysis 
on keyword and other metrics to discriminate between spam & nonspam. 
(It's the only way to be sure).

</$.02>

Tim