DUH!

Ok, I’ve gotta be like the dumbest person ever to not think of this before:

So PageRank (and the derivative that Google uses now) is susceptible to comment spamming in blogs because of the trust relationship that is created between the domain serving content and the linked-to domain. Blog comment spam that targets Google abuses this by re-weighting the links into the “target” domain and giving it un-earned weight (leaving the more general debate about the “earned” weight of blogs aside for now).

So here’s the thing: take away the incentive, take away the relationship, load all comments from a “sacrificial” domain and make sure any reference between the “loader” domain and the comment serving domain is tenuous at best (perhaps with JS or some other google-bot unfriendly mechanism). On this point, I’m not sure whether or not Simon’s redirect trick is essentially similar, but it would be a requirement.

The point here is that we don’t really want to discourage the spammers from posting (we’ll deal with that through baysean filtering or some other means), but we want to remove at least this incentive to spam.

On the spam-for-other-reasons front, I just mentioned baysean filtering, and I think that’s what is going to have to “win” (or something else that can adapt quickly, think spamassassin). To this end, running comments from a central service that can host the baysean/regex engine, serve you your comments as an IMAP folder (for integration with Thunderbird), and otherwise manage all aspects of your comment problems seems to make a lot of sense to me.

So what says the blog world at large? Would you pay a couple of bucks a months for this? or would you allow someone to put google ads next to your comment system in remuneration for hosting it? Or does something like this already exist and I missed it?