On Fri, 2009-10-23 at 08:55 +0100, Peter Ford wrote: > Ashley Sheridan wrote: > > > > > > Won't stop a bot worth it's salt either, hence the need for more complex > > and confusing captchas. The best way to stop spam, is to use linguistic > > testing on the content being offered, which protects against bot and > > human spammer alike. > > > > Thanks, > > Ash > > http://www.ashleysheridan.co.uk > > > > > > > > Unfortunately, it might also confound someone who doesn't speak the language. > Admittedly, they would probably already be struggling with the rest of the site... > > I guess locale-dependent captchas are a possibility. > > > -- > Peter Ford phone: 01580 893333 > Developer fax: 01580 893399 > Justcroft International Ltd., Staplehurst, Kent > I'm not talking about language problems for the user to solve. This question originally started by the op asking for solutions to human spam, but most of what I've seen so far in the thread is all about how to stop bots. By Linguistic analysis, I'm talking about passing the user offered content through a filter to check for the probability that it is spam. This goes beyond just looking for spammy words by looking at the relationship between words, frequency of words, and much more. It's very complex, but by the end of it, each post gets a rating value, which can be used as part of a threshold to reach in order to have a post automatically verified. Thanks, Ash http://www.ashleysheridan.co.uk
