[ckan-dev] Fwd: [ckan-discuss] CKAN spam

Tue Aug 2 11:11:06 UTC 2011

Adria,

I deleted and purged the offending revision, but unfortunately it
seems to have purged the whole package. I'm restoring it comparing
with the backup now.

I agree that the spam tools could be improved to cover this sort of
situation. I've added a ticket and copied my ideas below.
http://trac.ckan.org/ticket/1257 Please do make comments on this list
or on the ticket.

David

We are getting more and more spam on ckan.net and we need to improve
our strategy of combating it. It is bad because google ranks who we
link to (which we do want for legitimate links), and our front page
contains the latest package edits, so spam is immediately very
visible.

Spam:
 * Package creation
 * Package edit
 * User creation

Systems to consider:
 * Automatic
   + captcha
   + bayesian scoring
 * Sysadmin
   + a tool to quickly analyse and remove packages, edits and users
 * Helpful users
   + button to click indicating spam, leading to any/all of:
     + add TODO item (which emails package admin and sysadmins)
     + quarantine the package/user so it doesn't show up on the front
page and a warning is shown when you view it

General thoughts:
 * We should learn from the wikipedia tools, as they are the experts
 * This is a serious problem that is only going to get worse and is a
threat to the whole site.

David

2011/8/2 Adrià Mercader <amercadero at gmail.com>:
> Hi,
> I'm forwarding this to the dev list because I think on this occasion
> the issue is more severe than a regular spam package.
> I was flagging this package with meta.spam:
> http://ckan.net/package/alexb327
>
> But I checked the history and looks like it was not created as spam,
> but rather an existing legitimate package has been edited:
> http://ckan.net/package/history/alexb327
>
> If that's the case, I think that it's a thing that must be fixed asap
> to prevent losing any hard work made by the contributors.
>
>
> Adrià
>
>
> ---------- Forwarded message ----------
> From: Lucy Chambers <lucy.chambers at okfn.org>
> Date: 2011/8/2
> Subject: Re: [ckan-discuss] CKAN spam
> To: Bill Roberts <bill at swirrl.com>
> Cc: CKAN discuss <ckan-discuss at lists.okfn.org>
>
>
> Hi Bill,
>
> Thanks very much. Cc'ing CKAN-discuss.
>
> @Team, In history I can only compare versions, I don't appear to
> actually be able to roll back with my level of access, otherwise I
> would do this myself.
>
> L
>
> On Tue, Aug 2, 2011 at 9:09 AM, Bill Roberts <bill at swirrl.com> wrote:
>> Hi Lucy
>> Not sure who is responsible for maintaining CKAN, so perhaps you could
>> forward this to the right person for me.
>> I noticed a couple of recent spam packages have been added to ckan.net and
>> appear at the top of the front page (as the newest entries). I don't seem to
>> have the rights to delete them myself but perhaps someone could do some
>> tidying up.
>> Best regards
>> Bill
>
>
>
> --
> Lucy Chambers
> Community Coordinator
> Open Knowledge Foundation
> http://okfn.org/
> Skype: lucyfediachambers
>
> _______________________________________________
> ckan-discuss mailing list
> ckan-discuss at lists.okfn.org
> http://lists.okfn.org/mailman/listinfo/ckan-discuss
>
> _______________________________________________
> ckan-dev mailing list
> ckan-dev at lists.okfn.org
> http://lists.okfn.org/mailman/listinfo/ckan-dev
>