I think that your question is confusing people between anonymous “sources” (who tip off reporters to the fact that “there is a story to be told here, and here is something you need to investigate and write about”) with anonymous “commenters” (who comment on stories posted online, either with made up usernames – as we do in Fluther – or with completely anonymous “guest” tags or the like).
I believe that your question relates to the latter.
My own feeling is that if a news organization is going to operate a website that allows or encourages comments to be posted there, then it should exercise some minimum level of moderation and on its own kill comments that are simply posted as outright lies, flame bait, outright slander, or other forms of inflammatory or similar posts that detract from the “community” that the organization is attempting to establish by allowing and maintaining the forum in the first place.
In case that’s confusing: If it wouldn’t pass muster as public conversation at the water cooler in the newsroom, then it shouldn’t be kept on the website, either. There is nothing wrong with “robust” conversation, even opinions that we consider to be “extreme” or “misguided”, but commentary that is posted maliciously (or seems to be) doesn’t have to be allowed to stand.
On the other hand, if someone boasts of having committed a serious crime (or intending to commit or promote one), then why should those persons have any protection at all? They aren’t “sources” that the news organization seeks to keep private as information generators (a whole other topic of discussion, and which has already been much discussed in all forms of media – and in the courts).