CKAN is getting more 'backhanded compliments' in the form of spamming: http://www.ckan.net/revision/ http://www.ckan.net/revision/read/1978
Admins can keep purging this by hand [1] but this isn't hugely efficient. In fact around christmas to deal with a bad repeated attack I implemented some more sophisticated support ([2],[3]) including blacklisting but didn't fully integrate this into controllers. Before we go ahead and do more I wonder if anyone else has comments as to how best to deal with spam on 'world-editable' systems. In particular what are people's views on the effectiveness (and cost of implementation) of things like: * captchas (and texchas) * ip blacklisting * bayesian spam filtering of some kind ~rufus [1]:<http://lists.okfn.org/pipermail/okfn-help/2007-October/000038.html> [2]:<http://knowledgeforge.net/ckan/trac/changeset/205> [3]:<http://knowledgeforge.net/ckan/trac/changeset/202> _______________________________________________ okfn-discuss mailing list [email protected] http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss
