Excellent project idea. This punches right into epistemology - which is how we know what we know. It'd be great to either be quite opinionated about your epistemology, moreso than the article above outlines, or be a bit more agnostic about it and allow users to set things.
A big part of consensus forming is in communities of discourse - that is what that community talks about, what they assume to be true and how that community tests new information. This matter because different disciplines have different ways of doing that and for good reasons. For example, the scientific method is great for many things to do with the observable world, is somewhat okay for social questions and terrible for philosophical debates.
There's already some work on rating knowledge in the semantic web community and it might be a good starting point to work on from there. I wonder if @grampo wants to weigh in?