The War on Comment Spam

[Andy]

I found an interesting site by a MS Research team who are studying ways to eliminate comment spam on forums and blogs by tracing the source of spam posts.

http://research.microsoft.com/SearchDefender/

You can see that all the spam is due to the need to get inbound links to sites to get high search engine rankings. So I wonder if they have the correct approach to the problem?

It is most easy to write a search engine ranking algorithm to evaluate a sites worth by evaluating inbound links since these are stored in the search engine database. But it is difficult to find new websites that are full of new original content.

In my opinion, what is needed is a very good content analysis algorithm that can tell if your site contains good original content. Blogs stated out like this, so search engines found an easy way to find good new content to list. Then spammers made automated splogging tools to exploit this medium.

Now it seems that search engines are still taking the easy route to list sites e.g. based on age, dmoz listing etc. So the minority of good new sites are likely to go unnoticed for years in a sea of auto-generated spam sites.

[SensoVision]

I didn't read document throughly, but it seems that they provide a tool to analyze site for already posted SPAM and not intercept it during posting or maybe I misunderstood something?
Personally I believe that this problem should be addressed either in co-operation of domains and hosting companies and registrars or just search engine algorithms which could analyze target resource for it's content and rank it respectively.

I've also got another idea which could possibly be effective. Maybe it's worth to make some plug-in for Firefox which would let people ban sites like this and if it receive enough votes it could be checked by person and added to ban, maybe checking by human isn't worth... I didn't have this idea formed completely yet.