Carlton B wrote: the basic filter based on the spam lenght made me rid of this one (given to me in a previous post, a month ago) jdd -- http://www.dodin.net http://dodin.org/galerie_photo_web/expo/index.html http://lucien.dodin.net http://fr.susewiki.org/index.php?title=Gérer_ses_photos

On Sat, May 06, 2006 at 12:05:18PM -0400, Carlton B wrote: Different spammers and spambots use different strategies. There is no silver bullet solution but my strategy which seems to work very well for me - less than one spam attack per month in recent months - is a combination of methods: o Require users to create accounts. This is very useful anyway, you want to be able to contact authors occasionaly the one or other reason. o Use SORBS. Unlike email where SORBS being a rather big calibre of black list frequently causes false positives, SORBS has not triggered any such false positive for me yet. o Use $wgSpamRegex to catch typical spam patterns. In my case this were free hosters, attempts to use 1 pixel height fonts or other "stealth" techniques. o Reduce the value of spamming the wiki by setting $wgNoFollowLinks to true. This is the default since like 1.3, keep it. o I distructivly remove any spam from the database - as if it never happened. On my wishlist: some content analysis systems to augment or replace $wgSpamRegex. I was thinking of something along the lines of spamassassin, possibly even recycling some of it's code. Ralf